U.S. patent application number 14/199793 was filed with the patent office on 2014-10-09 for precoding matrix codebook design for advanced wireless communications systems.
This patent application is currently assigned to Samsung Electronics Co., Ltd.. The applicant listed for this patent is Samsung Electronics Co., Ltd.. Invention is credited to Young-Han Nam, Md Saifur Rahman, Yan Xin, Jianzhong Zhang.
Application Number | 20140301492 14/199793 |
Document ID | / |
Family ID | 51492077 |
Filed Date | 2014-10-09 |
United States Patent
Application |
20140301492 |
Kind Code |
A1 |
Xin; Yan ; et al. |
October 9, 2014 |
PRECODING MATRIX CODEBOOK DESIGN FOR ADVANCED WIRELESS
COMMUNICATIONS SYSTEMS
Abstract
Methods and apparatus of constructing rank-1 and/or rank-2
codebooks for advanced communication systems with 4 transmit
antennas and two-dimensional (2D) M.times.N transmit antenna
elements are provided. A double-codebook structure is considered
for 4-Tx antenna configuration. Single-codebook and double-codebook
structures are considered for two-dimensional (2D) M.times.N
transmit antenna elements.
Inventors: |
Xin; Yan; (Princeton,
NJ) ; Nam; Young-Han; (Plano, TX) ; Rahman; Md
Saifur; (Richardson, TX) ; Zhang; Jianzhong;
(Plano, TX) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Samsung Electronics Co., Ltd. |
Suwon-si |
|
KR |
|
|
Assignee: |
Samsung Electronics Co.,
Ltd.
Suwon-si
KR
|
Family ID: |
51492077 |
Appl. No.: |
14/199793 |
Filed: |
March 6, 2014 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
61775231 |
Mar 8, 2013 |
|
|
|
61816416 |
Apr 26, 2013 |
|
|
|
61832632 |
Jun 7, 2013 |
|
|
|
61834269 |
Jun 12, 2013 |
|
|
|
61930887 |
Jan 23, 2014 |
|
|
|
Current U.S.
Class: |
375/267 |
Current CPC
Class: |
H04B 7/0639 20130101;
H04B 7/0456 20130101 |
Class at
Publication: |
375/267 |
International
Class: |
H04B 7/04 20060101
H04B007/04 |
Claims
1. A method comprising: transmitting, by a base station (BS) via
antenna array, a plurality of signals to at least one user
equipment (UE); and applying a codebook to the plurality of signals
prior to transmitting, wherein the codebook is designed with
one-dimensional 4-Tx and two-dimensional M.times.N antenna
elements.
2. The method of claim 1, wherein transmitting comprising
transmitting via a two dimensional (2D) antenna array comprising a
number M of antenna elements and N antenna ports configured in a 2D
grid NH.times.NV and wherein the codebook comprises a double
codebook W.sub.1W.sub.2.
3. The method of claim 1, wherein transmitting comprising
transmitting via a two dimensional (2D) antenna array comprising a
number M of antenna elements and N antenna ports configured in a 2D
grid NH.times.NV and wherein the codebook comprises a single
codebook W.
4. The method of claim 1, wherein transmitting comprising
transmitting via a two dimensional (2D) antenna array comprising a
number M of antenna elements and N antenna ports configured in a 2D
grid NH.times.NV and wherein the codebook comprises a rank 1 double
codebook W.sub.1W.sub.2.
5. The method of claim 4, wherein a rank 1 codebook comprises
non-evenly oversampled DFT vectors.
6. The method of claim 4, wherein the rank 1 double codebook
includes the vectors that are a Kronecker product of the horizontal
steering vectors and the vertical steering vectors by uniformly
quantizing the functions of elevation and azimuth angles.
7. The method of claim 6, wherein the rank 1 double codebook
includes the Kronecker product of the horizontal and vertical
steering vectors constructed by non-uniformly quantizing the
functions of elevation and azimuth angles.
8. The method of claim 6, wherein the rank 1 double codebook
includes at least one of: the horizontal and elevation steering
vectors that use the dependency between elevation and azimuth
angles; the horizontal and elevation steering vectors constructed
by uniformly quantizing elevation and azimuth angles; and the outer
codebook W.sub.2 based on co-phasing factors depending on the beam
selection index and the number of times of beams appearing in
codewords.
9. The method of claim 6, wherein the rank 1 double codebook is
configurable for low-rise and high-rise scenarios.
10. For use in a wireless communication network, a base station
comprising: a two dimensional (2D) antenna array comprising a
number M of antenna elements and N antenna ports configured in a 2D
grid NH.times.NV, the 2D antenna array configured to communicate
with at least one subscriber station; and a controller configured
to apply a codebook to the signal, wherein the codebook is designed
with one-dimensional 4-Tx and two-dimensional M.times.N antenna
elements.
11. The base station of claim 10, wherein the 2D antenna array
comprises a 4-Tx antenna array and wherein the codebook comprises a
double codebook W.sub.1W.sub.2 with one-dimensional 4-Tx is
applied, the double codebook comprising an inner codebook W.sub.1
and an outer codebook W.sub.2.
12. The base station of claim 10, wherein the 2D antenna array
comprises a number M of antenna elements and N antenna ports
configured in a 2D grid NH.times.NV and wherein the codebook
comprises a double codebook W.sub.1W.sub.2.
13. The base station of claim 10, wherein the 2D antenna array
comprises a number M of antenna elements and N antenna ports
configured in a 2D grid NH.times.NV and wherein the codebook
comprises a single codebook W.
14. The base station of claim 10, wherein the 2D antenna array
comprises a number M of antenna elements and N antenna ports
configured in a 2D grid NH.times.NV and wherein the codebook
comprises a rank 1 double codebook W.sub.1W.sub.2.
15. The base station of claim 14, wherein a rank 1 codebook
comprises non-evenly oversampled DFT vectors.
16. The base station of claim 14, wherein the rank 1 double
codebook includes the vectors that are a Kronecker product of the
horizontal steering vectors and the vertical steering vectors by
uniformly quantizing the functions of elevation and azimuth
angles.
17. The base station of claim 16, wherein the rank 1 double
codebook includes the Kronecker product of the horizontal and
vertical steering vectors constructed by non-uniformly quantizing
the functions of elevation and azimuth angles.
18. The base station of claim 16, wherein the rank 1 double
codebook includes at least one of: the horizontal and elevation
steering vectors that use the dependency between elevation and
azimuth angles; and the horizontal and elevation steering vectors
constructed by uniformly quantizing elevation and azimuth
angles.
19. The base station of claim 16, wherein the rank 1 double
codebook is configurable for low-rise and high-rise scenarios.
20. The base station of claim 16, wherein the rank 1 double
codebook includes the outer codebook W.sub.2 based on co-phasing
factors depending on the beam selection index and the number of
times of beams appearing in codewords.
Description
CROSS-REFERENCE TO RELATED APPLICATION(S) AND CLAIM OF PRIORITY
[0001] The present application claims priority to U.S. Provisional
Patent Application Ser. No. 61/775,231 filed Mar. 8, 2013, entitled
"Downlink MIMO Codebook Design for Advanced Wireless Communications
Systems," U.S. Provisional Patent Application Ser. No. 61/816,416,
filed Apr. 26, 2013, entitled "Downlink MIMO Codebook Design for
Advanced Wireless Communications Systems," U.S. Provisional Patent
Application Ser. No. 61/832,632, filed Jun. 7, 2013, entitled
"Downlink MIMO Codebook Design for Advanced Wireless Communications
Systems," U.S. Provisional Patent Application 61/834,269, filed
Jun. 12, 2013, entitled Downlink MIMO Codebook Design for Advanced
Wireless Communications Systems," and U.S. Provisional Patent
Application Ser. No. 61/930,887, filed Jan. 23, 2014, entitled
"Precoding Matrix Codebook Design for Advanced Wireless
Communications Systems". The content of the above-identified patent
documents are incorporated herein by reference.
TECHNICAL FIELD
[0002] The present application relates generally to wireless
communication systems and, more specifically, to scheduling of data
transmissions.
BACKGROUND
[0003] A communication system includes a DownLink (DL) that conveys
signals from transmission points such as Base Stations (BSs) or
NodeBs to User Equipments (UEs) and an UpLink (UL) that conveys
signals from UEs to reception points such as NodeBs. A UE, also
commonly referred to as a terminal or a mobile station, may be
fixed or mobile and may be a cellular phone, a personal computer
device, and the like. A NodeB, which is generally a fixed station,
may also be referred to as an access point or other equivalent
terminology.
[0004] Existing 4-Tx codebook in Release 11 does not perform well
for a cross-polarized (XP) antenna setup, which is a commonly used
antenna setup in practice. The target of the enhanced 4-Tx codebook
design is to improve performance for narrowly-spaced and
widely-spaced cross-polarized antenna setups.
[0005] A two-dimensional (2-D) active antenna array can be used for
advanced communication systems such as full dimension multiple
input multiple output (FD-MIMO) systems. In a 2-D active antenna
array, antenna elements are placed in the vertical and horizontal
directions. Codebooks designed for 2-D active antenna array contain
two components corresponding to the horizontal and vertical
components of underlying channel models.
SUMMARY
[0006] This disclosure provides a system and method for performing
a precoding matrix codebook design for use in advanced
communications systems.
[0007] In a first embodiment, a method is provided. The method
includes transmitting, by a base station (BS) via antenna array, a
plurality of signals to at least one user equipment (UE). The
method also includes applying a codebook to the plurality of
signals prior to transmitting, wherein the codebook is designed
with one-dimensional 4-Tx and two-dimensional M.times.N antenna
elements.
[0008] In a second embodiment, a base station is provided. The base
station includes a two dimensional (2D) antenna array comprising a
number M of antenna elements and N antenna ports configured in a 2D
grid NH.times.NV, the 2D antenna array configured to communicate
with at least one subscriber station. The base station also
includes a controller configured to apply a codebook to the signal,
wherein the codebook is designed with one-dimensional 4-Tx and
two-dimensional M.times.N antenna elements.
[0009] Before undertaking the DETAILED DESCRIPTION below, it may be
advantageous to set forth definitions of certain words and phrases
used throughout this patent document. The term "couple" and its
derivatives refer to any direct or indirect communication between
two or more elements, whether or not those elements are in physical
contact with one another. The terms "transmit," "receive," and
"communicate," as well as derivatives thereof, encompass both
direct and indirect communication. The terms "include" and
"comprise," as well as derivatives thereof, mean inclusion without
limitation. The term "or" is inclusive, meaning and/or. The phrase
"associated with," as well as derivatives thereof, means to
include, be included within, interconnect with, contain, be
contained within, connect to or with, couple to or with, be
communicable with, cooperate with, interleave, juxtapose, be
proximate to, be bound to or with, have, have a property of, have a
relationship to or with, or the like. The term "controller" means
any device, system or part thereof that controls at least one
operation. Such a controller may be implemented in hardware or a
combination of hardware and software and/or firmware. The
functionality associated with any particular controller may be
centralized or distributed, whether locally or remotely. The phrase
"at least one of," when used with a list of items, means that
different combinations of one or more of the listed items may be
used, and only one item in the list may be needed. For example, "at
least one of: A, B, and C" includes any of the following
combinations: A, B, C, A and B, A and C, B and C, and A and B and
C.
BRIEF DESCRIPTION OF THE DRAWINGS
[0010] For a more complete understanding of the present disclosure
and its advantages, reference is now made to the following
description taken in conjunction with the accompanying drawings, in
which like reference numerals represent like parts:
[0011] FIG. 1 illustrates a wireless network according to
embodiments of the present disclosure;
[0012] FIG. 2A illustrates a high-level diagram of a wireless
transmit path according to embodiments of the present
disclosure;
[0013] FIG. 2B illustrates a high-level diagram of a wireless
receive path according to embodiments of the present
disclosure;
[0014] FIG. 3 illustrates a user equipment according to embodiments
of the present disclosure;
[0015] FIG. 4 illustrates 4-Tx cross-polarized antenna
configuration according to embodiments of the present
disclosure;
[0016] FIG. 5 illustrates 4H.times.4V cross-polarized antenna
configuration in FD-MEMO according to embodiments of the present
disclosure;
[0017] FIG. 6 illustrates beam gains for 4-Tx codebook design;
[0018] FIG. 7 illustrates 2-dimensional (2-D) uniform rectangular
array according to embodiments of the present disclosure;
[0019] FIG. 8A illustrates 1-dimensional (1-D) co-polarized antenna
array according to embodiments of the present disclosure;
[0020] FIG. 8B illustrates RF chain for each antenna port according
to embodiments of the present disclosure;
[0021] FIG. 9 illustrates digital precoding operation according to
embodiments of the present disclosure;
[0022] FIG. 10 illustrates 2-D cross-polarized antenna array
according to embodiments of the present disclosure;
[0023] FIG. 11 illustrates closed-loop MIMO systems according to
embodiments of the present disclosure;
[0024] FIG. 12 illustrates beam gains of 3-bit non-evenly
oversampled DFT vectors according to embodiments of the present
disclosure;
[0025] FIG. 13 illustrates beam gains of 3-bit evenly oversampled
DFT vectors according to embodiments of the present disclosure;
[0026] FIG. 14 illustrates beam gains of 4-bit non-evenly
oversampled 4-Tx DFT matrix according to embodiments of the present
disclosure;
[0027] FIGS. 15A and 15B illustrate a 2D co-polarized antenna array
according to embodiments of the present disclosure;
[0028] FIG. 16 illustrates a 2D cross-polarized antenna array
according to embodiments of the present disclosure;
[0029] FIG. 17 illustrates two co-polarized arrays constructed from
the 2D cross-polarized array according to embodiments of the
present disclosure;
[0030] FIG. 18 illustrates the definition of angle .alpha.
according to embodiments of the present disclosure;
[0031] FIG. 19 illustrates the relationship between angles
.alpha..sub.0 and .theta..sub.0 according to embodiments of the
present disclosure;
[0032] FIG. 20 illustrates the feasible region of .theta. and
.alpha. according to embodiments of the present disclosure;
[0033] FIG. 21 illustrates the dependency of the exponents v and h
when .phi..sub.0.epsilon.[-.pi.,.pi.] according to embodiments of
the present disclosure;
[0034] FIG. 22 illustrates a high-rise scenario according to
embodiments of the present disclosure; and
[0035] FIG. 23 illustrates quantization of cosine values of the
elevation angles according to embodiments of the present
disclosure.
DETAILED DESCRIPTION
[0036] FIGS. 1 through 23, discussed below, and the various
embodiments used to describe the principles of the present
disclosure in this patent document are by way of illustration only
and should not be construed in any way to limit the scope of the
disclosure. Those skilled in the art will understand that the
principles of the present disclosure may be implemented in any
suitably arranged cellular system.
[0037] The following documents and standards descriptions are
hereby incorporated into the present disclosure as if fully set
forth herein: 3GPP TS 36.211 v11.1.0, "E-UTRA, Physical channels
and modulation" (REF 1); 3GPP TS 36.212 v11.1.0, "E-UTRA,
Multiplexing and Channel coding" (REF 2); 3GPP TS 36.213 v11.1.0,
"E-UTRA, Physical Layer Procedures" (REF 3); R1-130554, "On CSI
feedback enhancements", Ericsson, ST-Ericsson, Malta, January 2013
(REF 4); and R1-130012, "Performance evaluation of 4TX MU-MIMO in
scenario A", Huawei, January 2013 (REF 5).
[0038] FIG. 1 illustrates a wireless network 100 according to one
embodiment of the present disclosure. The embodiment of wireless
network 100 illustrated in FIG. 1 is for illustration only. Other
embodiments of wireless network 100 could be used without departing
from the scope of this disclosure.
[0039] The wireless network 100 includes NodeB 101, NodeB 102, and
NodeB 103. NodeB 101 communicates with NodeB 102 and NodeB 103.
NodeB 101 also communicates with Internet protocol (IP) network
130, such as the Internet, a proprietary IP network, or other data
network.
[0040] Depending on the network type, other well-known terms may be
used instead of "NodeB", such as "eNodeB" or "eNB," "transmission
point" (TP), "base station" (BS), "access point" (AP), or "eNodeB"
(eNB). For the sake of convenience, the term NodeB shall be used
herein to refer to the network infrastructure components that
provide wireless access to remote terminals.
[0041] For the sake of convenience, the term "user equipment" or
"UE" is used herein to designate any remote wireless equipment that
wirelessly accesses a NodeB, whether the UE is a mobile device
(e.g., cell phone) or is normally considered a stationary device
(e.g., desktop personal computer, vending machine, etc.). Also,
depending on the network type, other well-known terms may be used
instead of "user equipment" or "UE," such as "mobile station,"
"subscriber station," "remote terminal," "wireless terminal," or
"user device." For the sake of convenience, the terms "user
equipment" and "UE" are used in this patent document to refer to
remote wireless equipment that wirelessly accesses a NodeB, whether
the UE is a mobile device (such as a mobile telephone or
smartphone) or is normally considered a stationary device (such as
a desktop computer or vending machine).
[0042] NodeB 102 provides wireless broadband access to network 130
to a first plurality of user equipments (UEs) within coverage area
120 of NodeB 102. The first plurality of UEs includes UE 111, which
may be located in a small business; UE 112, which may be located in
an enterprise; UE 113, which may be located in a WiFi hotspot; UE
114, which may be located in a first residence; UE 115, which may
be located in a second residence; and UE 116, which may be a mobile
device, such as a cell phone, a wireless laptop, a wireless PDA, or
the like. UEs 111-116 may be any wireless communication device,
such as, but not limited to, a mobile phone, mobile PDA and any
mobile station (MS). NodeB 103 provides wireless broadband access
to a second plurality of UEs within coverage area 125 of NodeB 103.
The second plurality of UEs includes UE 115 and UE 116. In some
embodiments, one or more of NodeBs 101-103 can communicate with
each other and with UEs 111-116 using 5G, LTE, LTE-A, WiMAX, or
other advanced wireless communication techniques as described in
embodiments of the present disclosure.
[0043] Dotted lines show the approximate extents of coverage areas
120 and 125, which are shown as approximately circular for the
purposes of illustration and explanation only. It should be clearly
understood that the coverage areas associated with base stations,
for example, coverage areas 120 and 125, may have other shapes,
including irregular shapes, depending upon the configuration of the
base stations and variations in the radio environment associated
with natural and man-made obstructions.
[0044] As described in more detail below, one or more of NodeB 102
and NodeB 103 includes processing circuitry, such as transmit
circuitry, configured to perform precoding matrix codebook design
for advanced wireless communications systems.
[0045] Although FIG. 1 depicts one example of a wireless network
100, various changes may be made to FIG. 1. For example, another
type of data network, such as a wired network, may be substituted
for wireless network 100. In a wired network, network terminals may
replace NodeBs 101-103 and UEs 111-116. Wired connections may
replace the wireless connections depicted in FIG. 1.
[0046] FIG. 2A is a high-level diagram of a wireless transmit path.
FIG. 2B is a high-level diagram of a wireless receive path. In
FIGS. 2A and 2B, the transmit path 200 may be implemented, e.g., in
NodeB 102 and the receive path 250 may be implemented, e.g., in a
UE, such as UE 116 of FIG. 1. It will be understood, however, that
the receive path 250 could be implemented in a NodeB (e.g., NodeB
102 of FIG. 1) and the transmit path 200 could be implemented in a
UE. In certain embodiments, transmit path 200 and receive path 250
are configured to perform methods for uplink control channel
multiplexing in beamformed cellular systems as described in
embodiments of the present disclosure. Each of the eNBs 101-103 can
include a processor, or processing circuitry, configured to perform
methods for precoding matrix codebook design for advanced wireless
communications systems, as described in embodiments of the present
disclosure.
[0047] Transmit path 200 comprises channel coding and modulation
block 205, serial-to-parallel (S-to-P) block 210, Size N Inverse
Fast Fourier Transform (IFFT) block 215, parallel-to-serial
(P-to-S) block 220, add cyclic prefix block 225, and up-converter
(UC) 230. Receive path 250 comprises down-converter (DC) 255,
remove cyclic prefix block 260, serial-to-parallel (S-to-P) block
265, Size N Fast Fourier Transform (FFT) block 270,
parallel-to-serial (P-to-S) block 275, and channel decoding and
demodulation block 280.
[0048] At least some of the components in FIGS. 2A and 2B may be
implemented in software while other components may be implemented
by configurable hardware (e.g., one or more processors) or a
mixture of software and configurable hardware. In particular, it is
noted that the FFT blocks and the IFFT blocks described in this
disclosure document may be implemented as configurable software
algorithms, where the value of Size N may be modified according to
the implementation.
[0049] Furthermore, although this disclosure is directed to an
embodiment that implements the Fast Fourier Transform and the
Inverse Fast Fourier Transform, this is by way of illustration only
and should not be construed to limit the scope of the disclosure.
It will be appreciated that in an alternate embodiment of the
disclosure, the Fast Fourier Transform functions and the Inverse
Fast Fourier Transform functions may easily be replaced by Discrete
Fourier Transform (DFT) functions and Inverse Discrete Fourier
Transform (IDFT) functions, respectively. It will be appreciated
that for DFT and IDFT functions, the value of the N variable may be
any integer number (i.e., 1, 2, 3, 4, etc.), while for IDFT and
IFFT functions, the value of the N variable may be any integer
number that is a power of two (i.e., 1, 2, 4, 8, 16, etc.).
[0050] In transmit path 200, channel coding and modulation block
205 receives a set of information bits, applies coding (e.g., turbo
coding) and modulates (e.g., Quadrature Phase Shift Keying (QPSK)
or Quadrature Amplitude Modulation (QAM)) the input bits to produce
a sequence of frequency-domain modulation symbols.
Serial-to-parallel block 210 converts (i.e., de-multiplexes) the
serial modulated symbols to parallel data to produce N parallel
symbol streams where N is the IFFT/FFT size used in NodeB 102 and
UE 116. Size N IFFT block 215 then performs an IFFT operation on
the N parallel symbol streams to produce time-domain output
signals. Parallel-to-serial block 220 converts (i.e., multiplexes)
the parallel time-domain output symbols from Size N IFFT block 215
to produce a serial time-domain signal. Add cyclic prefix block 225
then inserts a cyclic prefix to the time-domain signal. Finally,
up-converter 230 modulates (i.e., up-converts) the output of add
cyclic prefix block 225 to RF frequency for transmission via a
wireless channel. The signal may also be filtered at baseband
before conversion to RF frequency.
[0051] The transmitted RF signal arrives at UE 116 after passing
through the wireless channel and reverse operations to those at
NodeB 102 are performed. Down-converter 255 down-converts the
received signal to baseband frequency and remove cyclic prefix
block 260 removes the cyclic prefix to produce the serial
time-domain baseband signal. Serial-to-parallel block 265 converts
the time-domain baseband signal to parallel time domain signals.
Size N FFT block 270 then performs an FFT algorithm to produce N
parallel frequency-domain signals. Parallel-to-serial block 275
converts the parallel frequency-domain signals to a sequence of
modulated data symbols. Channel decoding and demodulation block 280
demodulates and then decodes the modulated symbols to recover the
original input data stream.
[0052] Each of NodeBs 101-103 may implement a transmit path that is
analogous to transmitting in the downlink to UEs 111-116 and may
implement a receive path that is analogous to receiving in the
uplink from UEs 111-116. Similarly, each one of UEs 111-116 may
implement a transmit path corresponding to the architecture for
transmitting in the uplink to NodeBs 101-103 and may implement a
receive path corresponding to the architecture for receiving in the
downlink from NodeBs 101-103.
[0053] Each of the components in FIGS. 2A and 2B can be implemented
using only hardware or using a combination of hardware and
software/firmware. As a particular example, at least some of the
components in FIGS. 2A and 2B may be implemented in software, while
other components may be implemented by configurable hardware or a
mixture of software and configurable hardware. For instance, the
FFT block 270 and the IFFT block 215 may be implemented as
configurable software algorithms, where the value of size N may be
modified according to the implementation.
[0054] Furthermore, although described as using FFT and IFFT, this
is by way of illustration only and should not be construed to limit
the scope of this disclosure. Other types of transforms, such as
Discrete Fourier Transform (DFT) and Inverse Discrete Fourier
Transform (IDFT) functions, could be used. It will be appreciated
that the value of the variable N may be any integer number (such as
1, 2, 3, 4, or the like) for DFT and IDFT functions, while the
value of the variable N may be any integer number that is a power
of two (such as 1, 2, 4, 8, 16, or the like) for FFT and IFFT
functions.
[0055] Although FIGS. 2A and 2B illustrate examples of wireless
transmit and receive paths, various changes may be made to FIGS. 2A
and 2B. For example, various components in FIGS. 2A and 2B could be
combined, further subdivided, or omitted and additional components
could be added according to particular needs. Also, FIGS. 2A and 2B
are meant to illustrate examples of the types of transmit and
receive paths that could be used in a wireless network. Any other
suitable architectures could be used to support wireless
communications in a wireless network.
[0056] FIG. 3 illustrates a UE according to embodiments of the
present disclosure. The embodiment of user equipment, such as UE
116, illustrated in FIG. 3 is for illustration only. Other
embodiments of the wireless subscriber station could be used
without departing from the scope of this disclosure. Although MS
116 is depicted by way of example, the description of FIG. 3 can
apply equally to any of UE 111, UE 112, UE 113, UE 114 and UE
115.
[0057] UE 116 comprises antenna 305, radio frequency (RF)
transceiver 310, transmit (TX) processing circuitry 315, microphone
320, and receive (RX) processing circuitry 325. UE 116 also
comprises speaker 330, main processor 340, input/output (I/O)
interface (IF) 345, keypad 350, display 355, and memory 360. Memory
360 further comprises basic operating system (OS) program 361 and a
plurality of applications 362.
[0058] Radio frequency (RF) transceiver 310 receives from antenna
305 an incoming RF signal transmitted by a NodeB of wireless
network 100. Radio frequency (RF) transceiver 310 down-converts the
incoming RF signal to produce an intermediate frequency (IF) or a
baseband signal. The IF or baseband signal is sent to receiver (RX)
processing circuitry 325 that produces a processed baseband signal
by filtering, decoding, and/or digitizing the baseband or IF
signal. Receiver (RX) processing circuitry 325 transmits the
processed baseband signal to speaker 330 (such as for voice data)
or to main processor 340 for further processing (such as for web
browsing data).
[0059] Transmitter (TX) processing circuitry 315 receives analog or
digital voice data from microphone 320 or other outgoing baseband
data (e.g., web data, e-mail, interactive video game data) from
main processor 340. Transmitter (TX) processing circuitry 315
encodes, multiplexes, and/or digitizes the outgoing baseband data
to produce a processed baseband or IF signal. Radio frequency (RF)
transceiver 310 receives the outgoing processed baseband or IF
signal from transmitter (TX) processing circuitry 315. Radio
frequency (RF) transceiver 310 up-converts the baseband or IF
signal to a radio frequency (RF) signal that is transmitted via
antenna 305.
[0060] In certain embodiments, main processor 340 is a
microprocessor or microcontroller. Memory 360 is coupled to main
processor 340. According to some embodiments of the present
disclosure, part of memory 360 comprises a random access memory
(RAM) and another part of memory 360 comprises a Flash memory,
which acts as a read-only memory (ROM).
[0061] Main processor 340 can be comprised of one or more
processors and executes basic operating system (OS) program 361
stored in memory 360 in order to control the overall operation of
wireless subscriber station 116. In one such operation, main
processor 340 controls the reception of forward channel signals and
the transmission of reverse channel signals by radio frequency (RF)
transceiver 310, receiver (RX) processing circuitry 325, and
transmitter (TX) processing circuitry 315, in accordance with
well-known principles. Main processor 340 can include processing
circuitry configured to allocate one or more resources. For example
Main processor 340 can include allocator processing circuitry
configured to allocate a unique carrier indicator and detector
processing circuitry configured to detect a PDCCH scheduling a
PDSCH reception of a PUSCH transmission in one of the C
carriers.
[0062] Main processor 340 is capable of executing other processes
and programs resident in memory 360, such as operations for
precoding matrix codebook design for advanced wireless
communications systems as described in embodiments of the present
disclosure. Main processor 340 can move data into or out of memory
360, as required by an executing process. In some embodiments, the
main processor 340 is configured to execute a plurality of
applications 362, such as applications for MU-MIMO communications,
including obtaining control channel elements of PDCCHs. Main
processor 340 can operate the plurality of applications 362 based
on OS program 361 or in response to a signal received from BS 102.
Main processor 340 is also coupled to I/O interface 345. I/O
interface 345 provides subscriber station 116 with the ability to
connect to other devices such as laptop computers and handheld
computers. I/O interface 345 is the communication path between
these accessories and main controller 340.
[0063] Main processor 340 is also coupled to keypad 350 and display
unit 355. The operator of subscriber station 116 uses keypad 350 to
enter data into subscriber station 116. Display 355 may be a liquid
crystal display capable of rendering text and/or at least limited
graphics from web sites. Alternate embodiments may use other types
of displays.
[0064] Although FIG. 3 illustrates one example of UE 116, various
changes may be made to FIG. 3. For example, various components in
FIG. 3 could be combined, further subdivided, or omitted and
additional components could be added according to particular needs.
As a particular example, the main processor 340 could be divided
into multiple processors, such as one or more central processing
units (CPUs) and one or more graphics processing units (GPUs).
Also, while FIG. 3 illustrates the UE 116 configured as a mobile
telephone or smartphone, UEs could be configured to operate as
other types of mobile or stationary devices.
[0065] A NodeB transmits a PDCCH in units referred to as Control
Channel Elements (CCEs). The NodeB, such as NodeB 102 or NodeB 103,
transmits one or more of multiple types of RS including a UE-Common
RS (CRS), a Channel State Information RS (CSI-RS), and a
DeModulation RS (DMRS). The CRS is transmitted over substantially a
DL system BandWidth (BW) and can be used by the UEs, such as UE
116, to demodulate data or control signals or to perform
measurements. The UE 116 can determine a number of NodeB antenna
ports from which a CRS is transmitted through a broadcast channel
transmitted from the NodeB. To reduce the CRS overhead, the NodeB
can transmit a CSI-RS with a smaller density in the time and/or
frequency domain than the CRS. The UE can determine the CSI-RS
transmission parameters through higher layer signaling from the
NodeB. The DMRS is transmitted only in the BW of a respective PDSCH
and a UE can use the DMRS to demodulate the information in the
PDSCH.
[0066] FIG. 4 illustrates a 4 transmit (4-Tx) cross polarized (XP)
antenna setup according to embodiments of the present disclosure.
The embodiment of the 4-Tx XP antenna setup 900 shown in FIG. 4 is
for illustration only. Other embodiments could be used without
departing from the scope of the present disclosure.
[0067] Embodiments of the 4-Tx XP antenna setup 900 are commonly
used antenna setup in practice. Embodiments of the 4-Tx XP antenna
setup 900 include two pairs of transmit antennas, each of which
comprises one antenna (910 or 930) with polarization direction +45
degree and the other antenna (920 or 940) with polarization
direction -45 degree. That is, a first pair of antenna includes a
first antenna 910 with polarization direction +45.degree. and a
second antenna 920 with polarization direction -45.degree.. In
addition, a second pair of antenna includes a third antenna 930
with polarization direction +45.degree. and a fourth antenna 940
with polarization direction -45.degree..
[0068] FIG. 5 illustrates a BS antenna array equipped with
4.times.4 cross-polarized active antenna elements according to
embodiments of the present disclosure. The embodiment of the
antenna array 1000 shown in FIG. 5 is for illustration only. Other
embodiments could be used without departing from the scope of the
present disclosure.
[0069] In the example, shown in FIG. 5, the antenna array 1000
includes antenna pairs arranged in rows 1010, 1020, 1030, 1040 and
columns 1050, 1060, 1070, 1080. Each column includes 4 pairs of
cross-polarized antenna elements and each row includes 4 pairs
cross-polarized antenna elements. In each pair of cross-polarized
antennas, a first antenna has a +45.degree. polarization while the
second antenna has -45.degree. polarization.
[0070] A communication system consists of a downlink (DL), where
signals are transmitted from base station (BS), NodeBs or
transmission point (TP) to user equipment (UE), and an uplink (UL),
where signals are transmitted from UE to BS or NodeB. For example,
in the DL, NodeB 102 can transmit signals to UE 116. In the UL, UE
116 can transmit signals to NodeB 102.
[0071] A precoding matrix W Release 10 8-Tx codebook can be written
as a product of two precoding matrices, called inner and outer
precoders, respectively. The inner precoder has a block-diagonal
structure and is used to capture the wide-band and long-term
channel properties while the outer precoder is used to capture the
frequency-selective and short-term channel properties.
Mathematically, the precoding matrix W can be expressed as
W=W.sub.1W.sub.2, where W.sub.1 and W.sub.2 denote the inner and
outer precoders, respectively. Such structure is also called double
codebook W.sub.1W.sub.2 structure. The inner precoder W.sub.1 has
the following block diagonal structure:
W 1 = [ X 0 0 X ] . ( 2 ) ##EQU00001##
A 4-Tx codebook design based on the 8-Tx double codebook
W.sub.1W.sub.2 design principle is proposed. In particular, the
inner precoder W.sub.1 follows the same structure as given in
equation (2), where X is a 2.times.1 vector chosen from the
following set:
X .di-elect cons. { [ 1 j 2 .pi. n 16 ] , n = 0 , 1 , , 15 } ( 3 )
##EQU00002##
for rank 1 and 2. For rank 1, the outer precoder W.sub.2 is chosen
from the following set:
W 2 .di-elect cons. { [ 1 1 ] , [ 1 - 1 ] , [ 1 j ] , [ 1 - j ] } (
4 ) ##EQU00003##
and for rank 2, the outer precoder W.sub.2 is chosen from the
following set:
W 2 .di-elect cons. { [ 1 1 1 - 1 ] , [ 1 1 j - j ] } . ( 5 )
##EQU00004##
For rank 1, the overhead for transmitting PMI of W.sub.1 and PMI of
W.sub.2 are 4 bits and 2 bits, respectively. For rank 2, the
overhead for transmitting PMI of W.sub.1 and PMI of W.sub.2 are 4
bits and 1-bit, respectively. FIG. 6 illustrates example beam gains
of 4-Tx codebook designs according to REF 5.
[0072] FIG. 7 illustrates an azimuth angle and an elevation angle
(also called the zenith angle) according to embodiments of the
present disclosure. The embodiment of the azimuth angle and the
elevation angle shown in FIG. 7 is for illustration only. Other
embodiments could be used without departing from the scope of the
present disclosure. More specifically, the azimuth angle .phi. is
defined as the angle between the positive x-axis direction and the
project vector of a straight line between NodeB 102 and UE 116 to
the xy plane while the elevation angle is defined as the angle
between the positive z-axis and a straight line between NodeB 102
and UE 116.
[0073] FIG. 8A illustrates 1-D co-polarized antenna array, where N
virtual antenna ports, i.e., antenna ports (APs) 0 to N-1 are
placed on a substantially horizontal axis according to embodiments
of the present disclosure. The embodiment of the 1-D co-polarized
antenna array 1300 shown in FIG. 8A is for illustration only. Other
embodiments could be used without departing from the scope of the
present disclosure. In the example shown in FIG. 8A, each of
antenna ports is constructed with a number of physical antennas,
e.g., 1305.0.
[0074] FIG. 8B illustrates an example in which each antenna port is
mapped to M antennas placed on a substantially vertical axis 1355
according to embodiments of the present disclosure. The embodiment
of the antenna port mapping 1350 shown in FIG. 8B is for
illustration only. Other embodiments could be used without
departing from the scope of the present disclosure. Each of the M
antennas comprising an antenna port is connected to a phase shifter
1360. For each antenna port, a power amplifier block 1365 is
connected right after a carrier modulation block 1370, the signal
outputs of which are split into these M phase shifters. Baseband
transmit antenna input for each antenna port of size N.sub.IFFT
goes through an OFDM digital chain 1375, comprising IFFT, P/S, add
CP and DAC.
[0075] FIG. 9 illustrates a digital precoding operation according
to embodiments of the present disclosure. The embodiment of the
digital precoding operation 1400 shown in FIG. 9 is for
illustration only. Other embodiments could be used without
departing from the scope of the present disclosure. In the example
shown in FIG. 9, J input signals with J.ltoreq.N 1405 are input
into digital precoding matrix 1410. The baseband Tx input signals
1420 for the N antenna ports are constructed by pre-multiplying a
(N.times.J) digital precoding matrix 1410 to J input signals with
J.ltoreq.N 1405.
[0076] In the present disclosure, f.sub.c is defined as the carrier
frequency and .lamda..sub.c as the corresponding wavelength of the
carrier frequency f.sub.c. A time domain baseband equivalent model
h(.tau.) observed at a receive antenna at a UE is given by
c ( .tau. ) = l = 0 L - 1 .alpha. l a H ( h l ) .delta. ( .tau. -
.tau. l ) j 2 .pi. f c .tau. l , ( 6 ) ##EQU00005##
where L is defined as the number of multiple paths, .alpha..sub.l
denotes the a complex number associated with the lth path,
.theta..sub.l is the azimuth angle of the lth path, a(h.sub.l) is a
M.times.1 steering vector defined as a(h.sub.1)=[1 e.sup.-jh.sup.l
. . . e.sup.-j(M-1)] with
h l := 2 .pi. d V .lamda. c cos .theta. l , ##EQU00006##
and .tau..sub.l is defined as propagation delay of the lth path.
When the steering vector a(h.sub.l) with
h l := 2 .pi. d V .lamda. c cos .theta. l ##EQU00007##
is applied to M elements comprising an antenna port, the signal
emitted from the antenna array will have a beam steered to a
direction of .theta..sub.l in the elevation domain. Accordingly,
the frequency domain representation for the model described in (1)
is given by:
H(f)=.SIGMA..sub.l=0.sup.L-1.alpha..sub.la.sup.H(h.sub.l)e.sup.-j2.pi.(f-
-f.sup.c.sup.).tau..sup.l. (7)
The spatial covariance matrix of the channel H(f) at a specific
frequency is defined as:
R.sub.H=E[|.alpha..sub.0|.sup.2][a(h.sub.0)a.sup.H(h.sub.0)].
(8)
Correspondingly, the most dominant eigenvector of R.sub.H is
c=.beta.a(h.sub.0) with .beta. being the power normalizing
factor.
[0077] FIG. 10 illustrates 2-D cross-polarized antenna array
according to embodiments of the present disclosure. The embodiment
of the 2-D cross-polarized antenna array 1500 shown in FIG. 10 is
for illustration only. Other embodiments could be used without
departing from the scope of the present disclosure.
[0078] In the LIE Releases 10 and 12, a 8-Tx codebook and a 4-Tx
enhanced codebook are respectively specified in the standards. The
8-Tx codebook in Release 10 and the 4-Tx enhanced codebook in
Release 12 were designed for cross-polarized antenna as illustrated
in FIG. 10. In the cross-polarized antenna systems, each column of
antenna elements includes M pairs of antennas, wherein each pair
comprises a first antenna having polarization direction is
+45.degree., and a second antenna having polarization direction is
-45.degree.. Those M antennas having the same polarization
direction in each column comprise an antenna port. For example, for
the leftmost column, AP 0 shown in FIG. 10, antenna port 1510
comprises those M antenna elements with +45.degree. polarization,
and AP N/2-1 shown in FIG. 10, antenna port 1515 comprises those M
antenna elements with -45.degree. polarization. Those M antenna
elements comprising an antenna port can be fed by an RF chain, such
as RF chain 1350 shown in FIGS. 8A and 8B. The codebook for this
cross-polarized antenna system is designed based upon a
double-codebook structure, where a code word (a precoding matrix) W
can be expressed as a product of two precoding matrices W.sub.1 and
W.sub.2, i.e., W=W.sub.1W.sub.2. More specifically, the inner
precoding matrix W.sub.1 is used to capture the long term and
wide-band channel characteristics and the outer precoding matrix
W.sub.2 is used to capture the short term and frequency selective
channel characteristics. The rank-1 W.sub.1 in the 8-Tx double
codebook in Release 10 and the 4-Tx enhanced double codebook
Release 12 comprises discrete Fourier transform (DFT) vectors (see
REF3), intending to align with the beam steering vector a(h.sub.0)
to be applied for the co-polarized elements (i.e., either of
+45.degree. polarized elements or -45.degree. polarized elements in
Error! Reference source not found.) in the azimuth domain, taking a
form of Dr `l` vectors. In addition, the Release 8 4-Tx codebook
also contains DFT vectors as codewords.
[0079] NodeB 102 transmits cell-specific reference signals (CRS) to
facilitate its serving UEs' demodulation of control signals,
estimation of channel state information (CSI), and demodulation of
data carried on physical downlink shared channels (PDSCH). NodeB
102 also can configure CSI reference signals (CSI-RS) to facilitate
UEs CSI estimation. Upon processing reference signals, either CRS
or CSI-RS, UE 116 estimates CSI, which comprises at least one of
precoding matrix indicator (PMI), rank indicator (RI), and channel
quality indicator (CQI). UE 116 then feeds back the estimated CSI
over physical uplink control channel (PUCCH) using Format 2/2a/2b
with a payload size up to 11 bits or over physical uplink shared
channel (PUSCH) without payload size limitation, dependent upon
eNB's (or higher-layer configuration). The PUCCH feedback is often
configured with fixed period, while the PUSCH feedback is
dynamically triggered by NodeB 102 via CSI triggering bit(s)
carried in a uplink downlink control information (DCI) transmitted
on physical downlink control channels (PDCCH).
[0080] FIG. 11 illustrates a close-loop MIMO system according to
embodiments of the present disclosure. The embodiment of the
close-loop MIMO system 1600 shown in FIG. 11 is for illustration
only. Other embodiments could be used without departing from the
scope of the present disclosure.
[0081] A transmitter 1610 applies a precoding matrix W to input
signals X.sub.1 X.sub.2 and transmits a signal to receiver 1620.
The receiver 1620 in the closed-loop MIMO system estimates the
channels, searches the best precoding matrix, and feedback PMI and
other information to the transmitter. It is well-known that the
channel state information (CSI) at transmitter 1610 can be used to
improve the performance of MIMO systems. In the frequency division
duplexing (FDD) systems, acquiring CSI at the transmitter 1610
requires feedback from the receiver 1620. In particular, PMI is an
important type of CSI, which is needed to be fed back to the
transmitter 1610 with good accuracy. Alternatively, accurate PMI
feedback requires large communication overhead, which will degrade
the system performance. Therefore, an appropriate precoding
codebook designs plays an important role in achieving desirable
performance and overhead tradeoff.
[0082] In certain embodiments, in the proposed 4-Tx codebook
design, the same double codebook W.sub.1W.sub.2 structure is used
to obtain a new 4-Tx codebook, e.g., W=W.sub.1W.sub.2. In existing
enhanced 4-Tx codebook designs (see REF 3 and REF 4), the beam
vectors in X are obtained by evenly oversampling 4-Tx DFT vectors,
i.e.
X .di-elect cons. { [ 1 j 2 .pi. n 16 ] , n = 0 , 1 , , 15 } .
##EQU00008##
Unlike existing designs in (see REF 3 and REF 4), the design of
beam vectors are obtained by oversampling 4-Tx DFT vectors
unevenly. For 3-bit W.sub.1 codebook, the following designs
[ X 0 0 X ] ##EQU00009##
are provided wherein the 8 possible values of X are chosen from
{ [ 1 j2.pi. n 16 ] , n = 0 , 1 , , 15 } . ##EQU00010##
One example construction of these 8 possible values of X are shown
below:
X .di-elect cons. { [ 1 1 ] , [ 1 j2.pi. 16 ] , [ 1 j2.pi.2 16 ] ,
[ 1 j2.pi.5 16 ] , [ 1 j2.pi.10 16 ] , [ 1 j2.pi.13 16 ] , [ 1
j2.pi.14 16 ] , [ 1 j2.pi.15 16 ] } ( 9 ) ##EQU00011##
Since there are 8 possible inner precoders, the overhead for the
feedback of W.sub.1 is 3-bits.
[0083] FIG. 12 illustrates beam gains with respect to the azimuth
angular spread according to embodiments of the present disclosure.
The embodiment of the beam gains shown in FIG. 12 is for
illustration only. Other embodiments could be used without
departing from the scope of the present disclosure. For example,
the beam gains with respect to the azimuth angular spread shown in
FIG. 12 are for the proposed 4-Tx codebook design above. The outer
precoders for rank 1 and rank 2 are selected as ones proposed in
REF 3. Another example construction of the 3-bit 4-Tx codebook is
based on evenly sampled DFT vectors as follows
X .di-elect cons. { [ 1 j2.pi. n 8 ] , n = 0 , 1 , , 7 } .
##EQU00012##
The proposed 3-bit codebook has the approximately same resolution
for UE 116 located within angular spread around spread around 90
degree as 4-bit codebook proposed in REF 3. As compared with FIG.
13, which illustrates evenly oversampled DFT vectors, beam vectors
in near 0 and 180 degrees (vectors represented by
{ [ 1 j2.pi. n 16 ] , n = 5 , 10 } ) ##EQU00013##
with lower density. If UEs are densely located within small angular
spread centering at 90 degree, embodiments of the proposed 3-bit
codebook yield a similar performance as the evenly oversampled
4-bit 4-Tx DFT codebook.
[0084] In certain embodiments, the codebook for X is subset of
{ [ 1 j2.pi. n 16 ] , n = 0 , 1 , , 15 } , ##EQU00014##
where the subset comprises 8 distinct vectors indexed by 8 distinct
n values. The subset includes two groups of distinct vectors, where
one group has N.sub.1 elements and the other group has N.sub.2
elements, N.sub.1+N.sub.2=8. The first group that includes N.sub.1
elements has floor(N.sub.1/2) (or alternatively ceiling(N.sub.1/2))
consecutive integers increasing from 0 for n, and
ceiling(N.sub.1/2) (or alternatively floor(N.sub.1/2)) consecutive
integers decreasing from 15 for n. The N.sub.2 elements included in
the second group are coarsely sampled numbers from the integer
numbers from floor(N.sub.1/2) to 16-ceiling(N.sub.i/2) (or
alternatively ceiling(N.sub.1/2) to 16-floor(N.sub.1/2)).
Embodiments of this method can be easily generalized for arbitrary
set of
{ [ 1 j2.pi. n N ] , n = 0 , 1 , , N - 1 } , ##EQU00015##
where N is 2.sup.k, and k is a positive integer.
[0085] In certain embodiments, for 4-bit 4-Tx codebook, the design
of beam vectors that are obtained by oversampling 4-Tx DFT vectors
unevenly. For 4-bit 4-Tx W.sub.1 codebook, the following
designs
W 1 = [ X 0 0 X ] ##EQU00016##
with X chosen from the following set, which is illustrated in FIG.
14:
X .di-elect cons. { [ 1 1 ] , [ 1 j2.pi. 32 ] , [ 1 j2.pi.2 32 ] ,
[ 1 j2.pi.3 32 ] , [ 1 j2.pi.4 32 ] , [ 1 j7.pi. 16 ] , [ 1 j10.pi.
16 ] , [ 1 j13.pi. 16 ] , [ 1 j16.pi. 16 ] , [ 1 j19.pi. 16 ] , [ 1
j22.pi. 16 ] , [ 1 j25.pi. 16 ] , [ 1 j2.pi.28 32 ] , [ 1 j2.pi.29
32 ] , [ 1 j2.pi.30 32 ] , [ 1 j2.pi.31 32 ] } . ( 10 )
##EQU00017##
[0086] Considering the oversampling factor is a power of 2, the
n-bit 4-Tx inner precoders are designed based on unevenly
oversampled DFT vectors. The inner precoder W.sub.1 has the same
block diagonal structure as Equation (1). The vector X is obtained
as follows, where
N = 2 n and .DELTA. = 3 N + 2 N + 2 . ##EQU00018##
The vector X is chosen from the following set:
X .di-elect cons. { [ 1 j2.pi. l 2 N ] , for l = 0 , , n 4 - 1 , or
[ 1 j2.pi. l 2 N ] , l = 7 N 4 , , 2 N - 1 , [ 1 j2.pi. 2 N ( ( N 4
- 1 ) + k .DELTA. ) ] , k = 1 , , N 2 } . ( 11 ) ##EQU00019##
In this embodiment, beam vectors are restricted to the set that
contains the beam vectors are oversampled by a factor of 2 to a
power. This construction leads denser beam distribution centering
at 90.degree. and sparser beam distribution near 0 and 180.degree..
The outer precoder W.sub.2 is selected as the same as ones in REF 3
for rank 1 and rank 2. Note that this design does not assume
Embodiment 1 and Embodiment 2 as special cases. For example, for
n=4, the vector X is chosen from the following set
X .di-elect cons. { [ 1 1 ] , [ 1 j2.pi. 32 ] , [ 1 j2.pi.2 32 ] ,
[ 1 j2.pi.3 32 ] , [ 1 j2.pi. ( 3 + .DELTA. ) 32 ] , [ 1 j2.pi. ( 3
+ 2 .DELTA. ) 32 ] , [ 1 j2.pi. ( 3 + 3 .DELTA. ) 32 ] , [ 1 j2.pi.
( 3 + 4 .DELTA. ) 32 ] , [ 1 j2.pi. ( 3 + 5 .DELTA. ) 32 ] , [ 1
j2.pi. ( 3 + 6 .DELTA. ) 32 ] , [ 1 j2.pi. ( 3 + 7 .DELTA. ) 32 ] ,
[ 1 j2.pi. ( 3 + 8 .DELTA. ) 32 ] , [ 1 j2.pi.28 32 ] , [ 1
j2.pi.29 32 ] , [ 1 j2.pi.30 32 ] , [ 1 j2.pi.29 32 ] } where
.DELTA. = 25 9 . ( 12 ) ##EQU00020##
[0087] Consider more general choices for the beam vector X as
compared with Embodiment 3. For the n-bit design, beam vectors X
are selected from the following set
X .di-elect cons. { [ 1 j2.pi. s ( n ) ] , for n = 0 , , N - 1 } ,
( 13 ) ##EQU00021##
Where s(n) denotes a real sequence of N elements. In particular,
the sequence s(n) can be a sequence of rational numbers, which can
be expressed as
s ( n ) = p n q n , n = 0 , , N - 1 , ##EQU00022##
with p.sub.n,q.sub.n being integers.
[0088] Consider the 4-Tx codebook designs for 4 bit W.sub.1 and 4
bit W.sub.2 for rank-1 and rank-2 subject to the following
restrictions:
[0089] 1) the number of beams N.sub.b in each W.sub.1 is restricted
to 2 or 4, and
[0090] 2) the precoding matrices W.sub.1 are the same for both rank
1 and rank 2.
The target antenna configurations for this embodiment are both
closely-spaced and widely-spaced cross-polarized 4-Tx antennas.
Taking into account performances in both closely-spaced and
widely-spaced cross-polarized 4-Tx antennas, each group of N.sub.b
beams consists of so called almost parallel beams (angle between 2
beams is close to 0 degree) and orthogonal beams (angle between 2
beams is 90 degree). Considering slow time-varying channel effects
and edge effects of sub-band frequency-selective precoding, 0, 1 or
2 overlapping beams are allowed between W.sub.k and W.sub.k+1. For
N.sub.b=4, express
W 1 as W 1 = [ X k 0 0 X k ] , k = 0 , 1 , 2 , , 15. ( 15 )
##EQU00023##
[0091] Design for 4 bit W.sub.1 with N.sub.b=4 and Define the
vector u.sub.k as:
u k = [ 1 q 1 k ] with q 1 - j2.pi. / Q 1 . ( 16 ) .
##EQU00024##
[0092] For Q.sub.1=16,
[0093] Option A1: In this option, the matrix X.sub.k is selected to
be:
X.sub.k={[u.sub.k mod 16u.sub.(k+1)mod 16u.sub.(k+2)mod
16u.sub.(k+9)mod 16]}, k=0, . . . , 15. (17)
In this design, there are two overleaping beams between X.sub.k and
X.sub.k+1. For each X.sub.k, a pair of orthogonal beams, i.e.,
u.sub.(k+1)mod 16 and u.sub.(k+9)mod 16; and three beams u.sub.k
mod 16 u.sub.(k+1)mod 16 u.sub.(k+2)mod 16 are almost parallel.
[0094] Option A2: In this option, the matrix X.sub.k is selected to
be:
X.sub.k={[u.sub.k mod 16u.sub.(k+1)mod 16u.sub.(k+8)mod
16u.sub.(k+9)mod 16]}; k=0, . . . , 15. (18).
In this design, there is a single overleaping beams between X.sub.k
and X.sub.k+1. For each X.sub.k, two pairs of orthogonal beams,
i.e., u.sub.k mod 16 and u.sub.(k+8)mod 16; and u.sub.(k+1)mod 16
and u.sub.(k+9)mod 16, and two beams u.sub.k mod 16 and
u.sub.(k+1)mod 16 are almost parallel.
[0095] For Q.sub.1=32
[0096] Option B1: In this option, the matrix X.sub.k is selected
as
X.sub.k={[u.sub.2k mod 32u.sub.(2k+1)mod 32u.sub.(2k+2)mod
32u.sub.(2k+17)mod 32]}; k=0, . . . , 15 (19).
In this design, there is a single overleaping beam between X.sub.k
and X.sub.k+1. For each X.sub.k, a pair of orthogonal beams, i.e.
u.sub.(2k+1)mod 32 and u.sub.(2k+17)mod 32, and three beams
u.sub.2k mod 32, u.sub.(2k+1)mod 32, u.sub.(2k+2)mod 32 are almost
parallel.
[0097] Option B2: In this option, the matrix X.sub.k is selected
as:
X.sub.k={[u.sub.2k mod 32u.sub.(2k+1)mod 32u.sub.(2k+16)mod
32u.sub.(2k+17)mod 32]}, k=0, . . . , 15
In this design, there is no overleaping beam between X.sub.k and
X.sub.k+1. For each X.sub.k, two pairs of orthogonal beams, i.e.,
u.sub.2k mod 32 and u.sub.(2k+16)mod 32, and u.sub.(2k+1)mod 32 and
u.sub.(2k+17)mod 32, and two beams u.sub.2k mod 32, and
u.sub.(2k+1)mod 32 are almost parallel.
[0098] Option B3: In this option, the matrix X.sub.k is selected
as:
X.sub.k={[u.sub.k mod 32u.sub.(k+8)mod 32u.sub.(k+16)mod
32u.sub.(k+24)mod 32]}), k=0, . . . , 7 (20).
X.sub.k={[v.sub.k mod 32v.sub.(k+8)mod 32v.sub.(k+16)mod
32v.sub.(k+24)mod 32]}, k=8, . . . , 15 (21).
where
v k = [ 1 j 2 .pi. ( k + 1 2 ) 32 ] . ( 22 ) ##EQU00025##
[0099] Option B4: In this option, the matrix X.sub.k is as:
X k = { [ 1 1 1 1 q 1 k / 2 q 1 k 2 + 8 q 1 k 2 + 16 q 1 k 2 + 24 ]
} , k = 0 , , 15. ( 23 ) . ##EQU00026##
[0100] Option B5: In this option, the matrix X.sub.k is as:
X k = { [ 1 1 1 1 q 1 k q 1 k + 8 q 1 k + 16 q 1 k + 24 ] } , k = 0
, , 7 ; and ( 24 ) X k = { [ 1 1 1 1 q 1 k + 1 / 2 q 1 k + 8 + 1 /
2 q 1 k + 16 + 1 / 2 q 1 k + 24 + 1 / 2 ] } , k = 8 , , 15. ( 25 )
. ##EQU00027##
[0101] Option B6: In this option, the matrix X.sub.k is:
X k = { [ 1 1 1 1 q 1 k q 1 k + 8 q 1 k + 16 q 1 k + 24 ] } , k = 0
, , 15. ( 26 ) ##EQU00028##
which is the same as one given in Solution 2a below. For rank-2,
the precoding matrix W.sub.2 is given by:
W 2 .di-elect cons. { 1 2 [ Y 1 Y 2 .alpha. ( i ) Y 1 - .alpha. ( l
) Y 2 ] , 1 2 [ Y 1 Y 2 j .alpha. ( i ) Y 1 - j .alpha. ( l ) Y 2 ]
} . ( 27 ) ##EQU00029##
where .alpha.(i)=q.sub.1.sup.2(i-1) with q.sub.1:=e.sup.j2.pi./32
for (Y.sub.1,Y.sub.2)=(e.sub.i,e.sub.l).
[0102] In RAN1#73 meeting, it was agreed that one of the following
codebook solutions, 2a and solution 2b, will be chosen:
[0103] Solution 2a:
W 1 = [ X n 0 0 X n ] ##EQU00030##
where n=0, 1, . . . , 15
X n = [ 1 1 1 1 q 1 n q 1 n + 8 q 1 n + 16 q 1 n + 24 ]
##EQU00031##
where q.sub.l=e.sup.j2.pi./32 [0104] For rank 1,
[0104] W 2 , n .di-elect cons. { 1 2 [ Y .alpha. ( i ) Y ] , 1 2 [
Y j.alpha. ( i ) Y ] , 1 2 [ Y - .alpha. ( i ) Y ] , 1 2 [ Y -
j.alpha. ( i ) Y ] } ##EQU00032## [0105] and Y.epsilon.{e.sub.1,
e.sub.2, e.sub.3, e.sub.4} and .alpha.(i)=q.sub.1.sup.2(i-1);
[0106] For rank 2,
[0106] W 2 , n .di-elect cons. { 1 2 [ Y 1 Y 2 Y 1 - Y 2 ] , 1 2 [
Y 1 Y 2 j Y 1 - j Y 2 ] } ##EQU00033## [0107] and
(Y.sub.1,Y.sub.2)=(e.sub.i,e.sub.k).epsilon.{(e.sub.1,e.sub.1),
(e.sub.2,e.sub.2), (e.sub.3,e.sub.3), (e.sub.4,e.sub.4),
(e.sub.1,e.sub.2), (e.sub.2,e.sub.3), (e.sub.1,e.sub.4),
(e.sub.2,e.sub.4)};
[0108] Solution 2b:
W 1 = [ X n 0 0 X n ] ##EQU00034##
where n=0, 1, . . . , 15
X n = [ 1 1 1 1 q 1 n q 1 n + 8 q 1 n + 16 q 1 n + 24 ]
##EQU00035##
where q.sub.1=e.sup.j2.pi./32 [0109] For rank 1,
[0109] W 2 , n .di-elect cons. { 1 2 [ Y .alpha. ( i ) Y ] , 1 2 [
Y j.alpha. ( i ) Y ] , 1 2 [ Y - .alpha. ( i ) Y ] , 1 2 [ Y -
j.alpha. ( i ) Y ] } ##EQU00036## [0110] and Y.epsilon.{e.sub.1,
e.sub.2, e.sub.3, e.sub.4} and .alpha.(i)=q.sub.1.sup.2(i-1);
[0111] For rank 2,
[0111] W 2 , n .di-elect cons. { 1 2 [ Y 1 Y 2 Y 1 Y 2 ] , 1 2 [ Y
1 Y 2 Y 1 - Y 2 ] , 1 2 [ Y 1 Y 2 - Y 1 Y 2 ] , 1 2 [ Y 1 Y 2 - Y 1
- Y 2 ] } ##EQU00037## ( Y 1 , Y 2 ) .di-elect cons. { ( e 2 , e 4
) } ##EQU00037.2## and ##EQU00037.3## W 2 , n .di-elect cons. { 1 2
[ Y 1 Y 2 Y 1 - Y 2 ] , 1 2 [ Y 1 Y 2 j Y 1 - j Y 2 ] }
##EQU00037.4## ( Y 1 , Y 2 ) .di-elect cons. { ( e 1 , e 1 ) , ( e
2 , e 2 ) , ( e 3 , e 3 ) , ( e 4 , e 4 ) } ##EQU00037.5## and
##EQU00037.6## W 2 , n .di-elect cons. { 1 2 [ Y 1 Y 2 Y 2 - Y 1 ]
} ##EQU00037.7## ( Y 1 , Y 2 ) .di-elect cons. { ( e 1 , e 3 ) , (
e 2 , e 4 ) , ( e 3 , e 1 ) , ( e 4 , e 2 ) } ##EQU00037.8##
[0112] In these codebook proposals, Solution 2a and Solution 2b,
X.sub.n and X.sub.n+8 contain the same set of beams, which results
in a certain large number of duplicated rank-2 codewords. To solve
this issue, Option B3-B5 are proposed by making all X.sub.n have
distinct sets of beams (and hence all W's are distinct) for n=0, 1,
. . . , 15. Option B6 is proposed by modifying rank-2 W.sub.2.
Using B3-B5 options, the new W.sub.1 for rank-1 and rank-2 are
distinct. Therefore, the overall codewords Ware all distinct. For
option B6, the argument is given as follows. For Solution 2a and
2b, X.sub.n+8=X.sub.nP.sub..pi., where P.sub..pi. is a permutation
matrix defined as:
[ 0 0 0 1 1 0 0 0 0 1 0 0 0 0 1 0 ] . ( 28 ) ##EQU00038##
For n=0, 1, . . . , 15,
W 2 .di-elect cons. { 1 2 [ Y 1 Y 2 .alpha. ( i ) Y 1 - .alpha. ( l
) Y 2 ] , 1 2 [ Y 1 Y 2 j .alpha. ( i ) Y 1 - j .alpha. ( l ) Y 2 ]
} . ( 29 ) . ##EQU00039##
Alternatively, the overall codewords W belong to:
( 32 ) ##EQU00040## ( .alpha. ( i ) Y i , .alpha. ( l ) Y l ) = { (
.alpha. ( 1 ) e 1 , .alpha. ( 1 ) e 1 ) , ( .alpha. ( 2 ) e 2 ,
.alpha. ( 2 ) e 2 ) , ( .alpha. ( 3 ) e 3 , .alpha. ( 3 ) e 3 ) , (
.alpha. ( 4 ) e 4 , .alpha. ( 4 ) e 4 ) , ( .alpha. ( 1 ) e 1 ,
.alpha. ( 2 ) e 2 ) , ( .alpha. ( 2 ) e 2 , .alpha. ( 3 ) e 3 ) , (
.alpha. ( 1 ) e 1 , .alpha. ( 4 ) e 4 ) , ( .alpha. ( 2 ) e 2 ,
.alpha. ( 4 ) e 4 ) } . ##EQU00040.2##
For n=0, 1, . . . , 7
W .di-elect cons. { 1 2 [ X n Y 1 X n Y 2 .alpha. ( i ) X n Y 1 -
.alpha. ( l ) X n Y 2 ] , 1 2 [ X n Y 1 X n Y 2 j.alpha. ( i ) X n
Y 1 - j.alpha. ( l ) X n Y 2 ] } , for n = 0 , 1 , , 7. and ( 30 )
W .di-elect cons. { 1 2 [ X n P .pi. Y 1 X n P .pi. Y 2 .alpha. ( i
) X n P .pi. Y 1 - .alpha. ( l ) X n P .pi. Y 2 ] , 1 2 [ X n P
.pi. Y 1 X n P .pi. Y 2 j.alpha. ( i ) X n P .pi. Y 1 - j.alpha. (
l ) X n P .pi. Y 2 ] } , for n = 8 , 9 , , 15. ( 31 )
##EQU00041##
while for n=8, 9, . . . , 15
( 33 ) . ( .alpha. ( i ) P .pi. Y i , .alpha. ( l ) P .pi. Y l ) =
{ ( .alpha. ( 1 ) e 2 , .alpha. ( 1 ) e 2 ) , ( .alpha. ( 2 ) e 3 ,
.alpha. ( 2 ) e 3 ) , ( .alpha. ( 3 ) e 4 , .alpha. ( 3 ) e 4 ) , (
.alpha. ( 4 ) e 1 , .alpha. ( 4 ) e 1 ) , ( .alpha. ( 1 ) e 2 ,
.alpha. ( 2 ) e 3 ) , ( .alpha. ( 2 ) e 3 , .alpha. ( 3 ) e 4 ) , (
.alpha. ( 1 ) e 2 , .alpha. ( 4 ) e 1 ) , ( .alpha. ( 2 ) e 3 ,
.alpha. ( 4 ) e 1 ) } . ##EQU00042##
As a result, (.alpha.(i)Y.sub.i,
.alpha.(l)Y.sub.l).noteq.(.alpha.(i)P.sub..pi.Y.sub.i,
.alpha.(l)P.sub..pi.Y.sub.l). Therefore, there are no duplicated
codewords for rank 2.
[0113] Designs for 4 bit W.sub.1 with N.sub.b=2
[0114] For Q.sub.1=16,
[0115] Option C1: In this option, the matrix X.sub.k is selected
as:
X.sub.k={[u.sub.k mod 16u.sub.(k+1)mod 16]}, k=0, . . . , 15.
(34).
[0116] In this design, there is a single overleaping beams between
X.sub.k and X.sub.k+1. For each X.sub.k, two beams u.sub.k mod 16
and u.sub.(k+1)mod 16 are almost parallel.
[0117] Option C2: In this option, the matrix X.sub.k is selected
as:
X.sub.k={[u.sub.k mod 16u.sub.(k+8)mod 16]}, k=0, . . . , 15.
(35).
[0118] In this design, no overleaping beam between X.sub.k and
X.sub.k+1 occurs. For each X.sub.k, there is one pair of orthogonal
beams, i.e., u.sub.k mod 16 and u.sub.(k+8)mod 16.
[0119] Designs for 4 bit W.sub.2
[0120] For rank 1 and rank 2, the precoding matrix W.sub.2 is given
by:
W.sub.2=[a.sub.1 . . . a.sub.R] (36).
where a.sub.r is defined as:
a r = [ q 2 m q 2 n ] and q 2 = j2.pi. / Q 2 . ( 37 )
##EQU00043##
[0121] Option a:
[0122] Rank 1: In this option, the precoding matrix W.sub.2 is
chosen to be 8-Tx W.sub.2 (the same as 8-Tx codebook)
W 2 .di-elect cons. { 1 2 [ Y Y ] , 1 2 [ Y j Y ] , 1 2 [ Y - Y ] ,
1 2 [ Y j Y ] } . ( 38 ) ##EQU00044##
where Y belongs to Y.epsilon.{{tilde over (e)}{tilde over
(e.sub.1)},{tilde over (e)}{tilde over (e.sub.2)},{tilde over
(e)}{tilde over (e.sub.3)},{tilde over (e)}{tilde over (e.sub.4)}}
with {tilde over (e)}{tilde over (e.sub.l)} denoting the ith column
of the identity matrix. In this option, the same beam is selected
for two different polarizations.
[0123] Rank 2: In this option, the precoding matrix is the same as
8-Tx W.sub.2
W 2 .di-elect cons. { 1 2 [ Y 1 Y 2 Y 1 - Y 2 ] , 1 2 [ Y 1 Y 2 j Y
1 - j Y 2 ] } . ( 39 ) ##EQU00045##
[0124] Option b:
[0125] Rank 1: In this option, the precoding matrix W.sub.2 is
chosen to be:
W 2 .di-elect cons. { 1 2 [ Y 1 Y 2 ] , 1 2 [ Y 1 j Y 2 ] , 1 2 [ Y
1 - Y 2 ] , 1 2 [ Y 1 j Y 2 ] } . ( 40 ) ##EQU00046##
where Y.sub.1, and Y.sub.2 belong to {{tilde over (e)}{tilde over
(e.sub.1)},{tilde over (e)}{tilde over (e.sub.2)},{tilde over
(e)}{tilde over (e.sub.3)},{tilde over (e)}{tilde over (e.sub.4)}}
with {tilde over (e)}{tilde over (e.sub.l)} denoting the ith column
of the identity matrix. In this option, the different beams can be
selected for two different polarizations.
[0126] Rank 2: In this option, the precoding matrix is the same as
8-Tx W.sub.2
W 2 .di-elect cons. { 1 2 [ Y 1 Y 2 Y 1 - Y 2 ] , 1 2 [ Y 1 Y 2 j Y
1 - j Y 2 ] } . ( 41 ) ##EQU00047##
[0127] For W.sub.1 given in options A1 and B1, one choice of the
pair (Y.sub.1,Y.sub.2) is given as follows:
(Y.sub.1,Y.sub.2).epsilon.{({tilde over (e.sub.1)},{tilde over
(e)}{tilde over (e.sub.1)}),({tilde over (e.sub.2)},{tilde over
(e)}{tilde over (e.sub.2)}),({tilde over (e.sub.3)},{tilde over
(e)}{tilde over (e.sub.3)}),({tilde over (e.sub.2)},{tilde over
(e)}{tilde over (e.sub.4)})} (42).
This choice enables the selection of two orthogonal beams (for
options A1 and B1, beam 2 and beam 4 are orthogonal) for two
different polarizations.
[0128] For W.sub.1 given in options A2 and B2, one choice of the
pair (Y.sub.1,Y.sub.2) is given as follows:
(Y.sub.1,Y.sub.2).epsilon.{({tilde over (e.sub.1)},{tilde over
(e)}{tilde over (e.sub.1)}),({tilde over (e.sub.2)},{tilde over
(e)}{tilde over (e.sub.2)}),({tilde over (e.sub.1)},{tilde over
(e)}{tilde over (e.sub.3)}),({tilde over (e.sub.2)},{tilde over
(e)}{tilde over (e.sub.4)})} (43).
This choice enables the selection of orthogonal beams (for options
A2 and B2, beam 1 and beam 3 are orthogonal and beam 2 and beam 4
are orthogonal) for two different polarizations.
[0129] Option c:
[0130] Rank 1: In Option c, the precoding matrix W.sub.2 is chosen
to be:
W 2 .di-elect cons. { 1 2 [ Y 1 q 2 m Y 2 ] } . ( 44 )
##EQU00048##
where Y.sub.1 and Y.sub.2 belong to {{tilde over (e)}{tilde over
(e.sub.1)}, {tilde over (e)}{tilde over (e.sub.2)}, {tilde over
(e)}{tilde over (e.sub.3)}, {tilde over (e)}{tilde over (e.sub.4)}}
with {tilde over (e)}{tilde over (e.sub.l)}(i=1, 2, 3, 4) denoting
the ith column of the identity matrix. In this option, the
different beams can be selected for two different polarizations.
One choice of Q.sub.2 is Q.sub.2=4 and m=0, 1, 2, 3. Hence,
q.sub.2.sup.m.epsilon.(1, j, -1, j). Generally, Q.sub.2=2.sup.N and
m.epsilon.a subset of {0, 1, . . . , Q.sub.2-1}.
[0131] Rank 2: In Option c, the precoding matrix is the same as
8-Tx W.sub.2
W 2 .di-elect cons. { 1 2 [ Y 1 Y 2 q 2 n Y 1 - q 2 n Y 2 ] , 1 2 [
Y 1 Y 1 q 2 n Y 2 - q 2 n Y 2 ] , 1 2 [ Y 1 Y 2 q 2 n Y 2 - q 1 k -
l q 2 n Y 1 ] } . ( 45 ) ##EQU00049##
where Y.sub.1 and Y.sub.2 belong to {{tilde over (e)}{tilde over
(e.sub.l)},{tilde over (e)}{tilde over (e.sub.k)}} and
k,l.epsilon.{1, 2, 3, 4}.
[0132] For W.sub.1 given in options A1 and B1, one choice of the
pair (Y.sub.1,Y.sub.2) is given as follows:
(Y.sub.1,Y.sub.2).epsilon.{({tilde over (e.sub.1)},{tilde over
(e)}{tilde over (e.sub.1)}),({tilde over (e.sub.2)},{tilde over
(e)}{tilde over (e.sub.2)}),({tilde over (e.sub.3)},{tilde over
(e)}{tilde over (e.sub.3)}),({tilde over (e.sub.2)},{tilde over
(e)}{tilde over (e.sub.4)})} (46).
This choice enables the selection of two orthogonal beams (for
options A1 and B1, beam 2 and beam 4 are orthogonal) for two
different polarizations.
[0133] For W.sub.1 given in options A2 and B2, one choice of the
pair (Y.sub.1,Y.sub.2) is given as follows:
(Y.sub.1,Y.sub.2).epsilon.{({tilde over (e.sub.1)},{tilde over
(e)}{tilde over (e.sub.1)}),({tilde over (e.sub.2)},{tilde over
(e)}{tilde over (e.sub.2)}),({tilde over (e.sub.1)},{tilde over
(e)}{tilde over (e.sub.3)}),({tilde over (e.sub.2)},{tilde over
(e)}{tilde over (e.sub.4)})} (47).
This choice enables the selection of orthogonal beams (for options
A2 and B2, beam 1 and beam 3 are orthogonal and beam 2 and beam 4
are orthogonal) for two different polarizations. The Kronecker
product based FD-MIMO codebook has a double codebook structure,
i.e., W=W.sub.1W.sub.2, where a codeword w.sub.1 in the W.sub.1
codebook can be expressed as w.sub.1=w.sub.Vw.sub.H, where w.sub.V
is a W.sub.1 codeword in the vertical domain called V-codebook and
w.sub.H is a W.sub.1 codeword in the elevation domain called
H-codebook. A codebook W for FD-MIMO with 8H.times.8V XP (64
antenna elements) antenna configuration is given as follows:
W V = { [ X V ( k ) 0 0 X V ( k ) ] , k = 0 , , 15 } and W H = { [
X H ( k ) 0 0 X H ( k ) ] , k = 0 , , 15 } , ( 48 ) .
##EQU00050##
where X.sub.V(k) and X.sub.H(k) are 4.times.4 matrices given by
X.sub.V(k)=[b.sub.f.sub.V,1.sub.(k)b.sub.f.sub.V,2.sub.(k)b.sub.f.sub.V,-
3.sub.(k)b.sub.f.sub.V,4.sub.(k)] (49).
X.sub.H(k)=[b.sub.f.sub.H,1.sub.(k)b.sub.f.sub.H,2.sub.(k)b.sub.f.sub.H,-
3.sub.(k)b.sub.f.sub.H,4.sub.(k)] (50).
with
b n = [ 1 j 2 .pi. n 32 j 2 .pi. 2 n 32 j 2 .pi. 3 n 32 ] T ,
##EQU00051##
n=0, 1, . . . , 31; and f.sub.V,t(k).epsilon.{0, 1, . . . , 31} and
f.sub.H,t(k).epsilon.{0, 1, . . . , 31}, k=1, 2, 3, 4. The Release
10 8-Tx codebook is used for the horizontal and vertical domains.
The overall FD-MIMO W.sub.1 codebook is given is constructed
as:
W 1 = { [ X ( i , j ) 0 0 X ( i , j ) ] , 0 .ltoreq. i .ltoreq. 15
, 0 .ltoreq. j .ltoreq. 15 } . ( 51 ) ##EQU00052##
with
b n = [ 1 j 2 .pi. n 32 j 2 .pi. 2 n 32 j 2 .pi. 3 n 32 ] T ,
##EQU00053##
n=0, 1, . . . , 31; and f.sub.V,t(k).epsilon.{0, 1, . . . , 31} and
f.sub.H,t(k).epsilon.{0, 1, . . . , 31}, k=1, 2, 3, 4. The Release
10 8-Tx codebook is used for the horizontal and vertical domains.
The overall FD-MIMO W.sub.1 codebook is given is constructed
as:
W 1 = { [ X ( i , j ) 0 0 X ( i , j ) ] , 0 .ltoreq. i .ltoreq. 15
, 0 .ltoreq. j .ltoreq. 15 } . ( 52 ) ##EQU00054##
For example: An evenly oversampled DFT codebook for 8V.times.8H
with 4-bit H-codebook and 4-bit V-codebook with
d.sub.H=d.sub.V=0.5.lamda..sub.c. .lamda..sub.V(k) and X.sub.H(k)
are 4.times.4 matrices given by:
[0134] X.sub.V(k)=X.sub.H(k)=[b.sub.2kmod 32b.sub.(2k+1)mod
32b.sub.(2k+2)mod32b.sub.(2k+3)mod32] with b.sub.n=
[ 1 j 2 .pi. n 32 j 2 .pi. 2 n 32 j 2 .pi. 3 n 32 ] T , n = 0 , 1 ,
, 31. ( 53 ) . ##EQU00055##
For example: A non-evenly oversampled DFT codebook for 8V.times.8H
with 4-bit H-codebook and 4-bit V-codebook with
d.sub.H=d.sub.V=0.5.lamda..sub.c.
[0135] The idea of using non-evenly oversampled DFT vector can be
applied to design the H-codebook and V-codebook. As stated earlier,
in the azimuth domain, UEs are located in a sector of 120.degree..
In the elevation domain, most UEs are distributed in an interval
of
[ .pi. 3 , 7 .pi. 12 ] . ##EQU00056##
A matrix X.sub.H (k) is defined as:
X.sub.H(k)=[b.sub.2kmod 32,Hb.sub.(2k+1)mod
32,Hb.sub.(2k+2)mod32,Hb.sub.(2k+3)mod32,H] (54).
for k=0, . . . , 15, where b.sub.n,H is defined as:
b n , H = [ 1 j 2 .pi. f ( n ) 32 j 2 .pi.2 f ( n ) 32 j 2 .pi.3 f
( n ) 32 ] T ##EQU00057##
for f(n)=0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 54, 55, 56, 57, 58, 59, 60,
61, 62, 63, respectively for n=0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 22,
23, 24, 25, 26, 27, 28, 29, 30, 31.
TABLE-US-00001 TABLE 1 The n-index versus f(n)-index n 0 1 2 3 4 5
6 7 8 9 22 23 24 25 26 27 28 29 30 31 f(n) 0 1 2 3 4 5 6 7 8 9 54
55 56 57 58 59 60 61 62 63
where 20 closely spaced beams constructed with 20 different values
of n correspond to azimuth steering angles around 90 degrees;
b n , H = [ 1 j 9 .pi. 32 + j 15 .pi. 128 ( n - 9 ) j 18 .pi. 32 +
j 15 .pi. 2 ( n - 9 ) 128 j 27 .pi. 32 + j 15 .pi. 3 ( n - 9 ) 128
] T ##EQU00058##
for n=10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21. where 12
widely spaced beams constructed with 12 different values of n
correspond to azimuth steering angles around 0 degrees. The matrix
X.sub.V(k) is defined as:
X.sub.V(k)=[b.sub.2kmod 32,Vb.sub.(2k+1)mod
32,Vb.sub.(2k+16)mod32,Vb.sub.(2k+17)mod32,V] for k=0, . . . ,
15.
This construction of X.sub.V(k) helps provide robust performance
for widely spaced and narrowly spaced antenna configuration, as it
contains two pairs of closely spaced beams and two pairs of
orthogonal beams. Widely spaced vertical antenna array can help to
separate beams in elevation domain even with small number of
antenna elements in the vertical antenna array, as it can provide
sufficient aperture.
b n , V = [ 1 j 2 .pi. fn 32 j 2 .pi.2 fn 32 j 2 .pi.3 fn 32 ] T
##EQU00059## for n.epsilon.{0, 1, . . . , 31} (55).
[0136] Certain embodiments show that an appropriate choice of a
co-phasing factor in the W2 codebook can reduce or eliminate
redundant codewords in the overall codebook. The Release 10 8-Tx
W.sub.1 codebook consists of 16 codewords, each containing 4
adjacent beams. There are two 2 overlapping beams between two
adjacent codewords. For rank 1, there are four beam selection
vectors and four co-phasing factors (i.e., 1, -1, j, -j), which are
constant independent of beam selection vectors. Due to the use of
the overlapping beams and beam selection independent co-phasing
factors, the final codebook W in Release 10 8-Tx codebook contains
a number of the redundant codewords for rank 1. In a FD-MIMO
codebook example shown in the present disclosure, each W.sub.1
codeword contains four overlapping beams with three other W.sub.1
codewords. To be more specific, since each pair of X(i) and X(i+1),
and X(j) and X(j+1) contains two overlapping beams, the FD-MIMO
codewords X(i,j), X(i+1,j), X(i,j+1), and X(i+1,j+1) have four
overlapping beams, which creates a large number of the redundant
codewords. This embodiment presents two methods for reducing or
eliminating the redundant codewords in the final codebook.
[0137] Method 1) reduce the number of the redundant beams in
X.sub.V(i) and/or X.sub.H(j) For rank 1-2, the W.sub.1 FD-MIMO
codebook for .alpha.V.times..beta.H for some positive even integers
.alpha. and .beta. can be constructed as
W 1 := { W 1 ( i , j ) = [ X ( i , j ) 0 0 X ( i , j ) ] , i = 0 ,
1 , , S v - 1 , j = 0 , 1 , , S h - 1 } . ( 56 ) ##EQU00060##
where X(i,j)=X.sub.V(i)*X.sub.H(j) with X.sub.V(i) and X.sub.H(j)
denoting the vertical and horizontal codebooks respectively. For
example, the matrix X.sub.V(i) can be selected as follows:
X.sub.V={X.sub.V(i)=[v.sub.(M.sub.1v.sub.i)modQ.sub.vv.sub.(M.sub.1v.sub-
.i+1)modQ.sub.v . . .
v.sub.(M.sub.1v.sub.i+M.sub.2v.sub.)modQ.sub.v]} (57).
where M.sub.1v, M.sub.2v and Q.sub.v are positive integers, and
v.sub.n is an oversampled vector given by
v n = [ 1 j 2 .pi. n Q v j 2 .pi. n ( .alpha. 2 - 1 ) Q v ] . ( 59
) ##EQU00061##
Similarly, the matrix X.sub.H (j) can be selected as follows:
X.sub.H={X.sub.H(j)=[h.sub.(M.sub.1h.sub.j)modQ.sub.s
hh.sub.(M.sub.1h.sub.j+1)modQ.sub.h . . .
h.sub.(M.sub.1h.sub.j+M.sub.2h.sub.)modQ.sub.h]}
where M.sub.1h, M.sub.2h and Q.sub.h are positive integers, and
h.sub.n is an oversampled DFT vector given by:
h n = [ 1 j 2 .pi. n Q h j 2 .pi. n ( .beta. 2 - 1 ) Q h ] . ( 60 )
##EQU00062##
Note that the number of the overlapping beams between X.sub.V(i)
and X.sub.V(i+1) is given by M.sub.2v-M.sub.1v+1, and the number of
the overlapping beams between X.sub.H(j) and X.sub.H(j+1) is given
by M.sub.2h-M.sub.1h+1. For instance, Release 10 8-Tx codebook has
M.sub.2h=3 and M.sub.1h=2, and thus there are two overlapping beams
between X.sub.H(j) and X.sub.H(j+1). Since the beams are normally
designed to cover the entire angular space [0,2.pi.], the
parameters M.sub.1v, M.sub.2v and Q.sub.v or M.sub.1h, M.sub.2h and
Q.sub.h, and the codebook size should be jointly designed. For
Q.sub.h=Q.sub.v=32, the choice of M.sub.2h=M.sub.2v=3 and M.sub.1h
M.sub.1v=2 leads to 4-bit X.sub.V an X.sub.H with two overlapping
beams while the choice of M.sub.2h=M.sub.2v=3 and
M.sub.1h=M.sub.1v=1 leads to 5-bit X.sub.V an X.sub.H with three
overlapping beams.
[0138] Method 2) Use co-phasing factor depending on the beam
selection indices in W.sub.2 In certain embodiments, the W.sub.2
codebook is designed as follows:
W 2 = { 1 2 [ Y .alpha. ( k ) Y ] } , where .alpha. ( k ) = exp ( j
2 .pi. k K ) , ( 61 ) . ##EQU00063##
k=0, . . . , K-1, and Y.epsilon.{e.sub.1, e.sub.2, . . . , e.sub.M}
with e.sub.j being defined as a 16.times.1 beam selection vector
with zero entries except for the ith entry. Clearly, the design of
the W.sub.2 codebook follows the same principle of the Release 10
8-Tx W.sub.2 codebook design. Note that in this proposed FD-MIMO
design, the choice of a(k) is independent of the beam selection
index.
[0139] Designing the co-phasing factors in W.sub.2 is another
approach to reduce or eliminate the redundant code-words. One
approach is to construct the W.sub.2 codebook as follows:
W 2 = { 1 2 [ Y .alpha. ( i ) Y ] , 1 2 [ Y j.alpha. ( i ) Y ] , 1
2 [ Y - .alpha. ( i ) Y ] , 1 2 [ Y - j.alpha. ( i ) Y ] } . ( 62 )
##EQU00064##
where .alpha.(i) depends on the beam selection index i. In the 4-Tx
enhanced codebook, a similar approach was used for designing the
rank-1 W.sub.2 codebook. Specifically, the choice of a(i) in the
4-Tx enhanced codebook is given by:
.alpha.(i)=e.sup.j(i-1).pi./8. (63).
It means that a beam selection index is determined as a function of
a co-phasing factor. In particular, in the 4-Tx enhanced codebook
case, the co-phasing factors are listed in the following table:
TABLE-US-00002 TABLE 2 Co-phasing factors versus beam selection
index i for the 4-Tx enhanced codebook beam index i 1 2 3 4 Co-
phasing factors {1, -1, j, -j} { e j .pi. 8 , - e j .pi. 8 ,
##EQU00065## e j 5 .pi. 8 , - e j 5 .pi. 8 } ##EQU00066## { e j
.pi. 4 , - e j .pi. 4 , ##EQU00067## e j 3 .pi. 4 , - e j 3 .pi. 4
} ##EQU00068## { e j 3 .pi. 8 , - e j 3 .pi. 8 , ##EQU00069## e j 7
.pi. 8 , - e j 7 .pi. 8 } ##EQU00070##
[0140] Referring to Table, the beam index in each code-word of the
W.sub.1 codebook determines the set of the co-phasing factors. The
set of the co-phasing factors {1, -1, j, -j} is the most commonly
used co-phasing factors and has been proved to have a better
codebook performance than other sets of co-phasing factors in
practice. In the 4-Tx enhanced codebook, the choice of the
co-phasing factors depends on the beam index. Depending upon the
W.sub.1 codebook design, this approach cannot guarantee that every
beam in the W.sub.1 codebook will use the set of the co-phasing
factors {1, -1, j, -j} since certain beams may not appear in every
column index in the W.sub.1 codebook. For example, in Release 10
8-Tx codebook design, the beam
b 3 = [ 1 j 2 .pi. 3 32 j 2 .pi.2 3 32 j 2 .pi. 3 3 32 ] T
##EQU00071##
only appears in column 4 of X(0) and appears in column 4 of X(1).
The set of the co-phasing factors {1, -1, j, -j} is not used to
apply the beam b.sub.3 if the construction approach for W.sub.2 in
the 4-Tx enhanced codebook is used. The following example shows a
different method to construct W.sub.2:
W 2 ( [ l 1 , n l 2 , n l 3 , n l 4 , n ] ) = { 1 2 [ e i .beta. (
l i , n ) e i ] , 1 2 [ e i j.beta. ( l i , n ) e i ] , 1 2 [ e i -
.beta. ( l i , n ) e i ] , 1 2 [ e i - j.beta. ( l i , n ) e i ] ,
i = 1 , 2 , 3 , 4 } . ( 64 ) ##EQU00072##
where .beta.(l.sub.i,n):=e.sup.j2.pi.l.sup.i,n.sup./8, and the
value of l.sub.i,n is the number of times of the appearance of the
ith beam in codeword W.sub.1(k) for k=0, 1, 2 . . . , n-1. Table
lists a W2 codebook design example. For example, in the first row
of Table 2 (n=0), all four beams b.sub.0 b.sub.1 b.sub.2 b.sub.3
have the set of co-phasing factors {1, -1, j, -j} corresponding to
l.sub.1,n=l.sub.2,n=l.sub.3,n=l.sub.4,n=0. In the second row of
Table 2 (n=1), the first and second beams b.sub.2, b.sub.3 have
appeared in W.sub.1(0) once and thus have the set of the co-phasing
factors
{ j.pi. 4 , - j.pi. 4 , j j.pi. 4 , - j j.pi. 4 } ##EQU00073##
corresponding to l.sub.1,n=l.sub.2,n=1, and the third and fourth
beams b.sub.4, b.sub.5 have not appeared in W.sub.1(0) and thus
have the set of the co-phasing factors {1, -1, j, -j} corresponding
to l.sub.3,n=l.sub.4,n=0.
TABLE-US-00003 TABLE 3 The W.sub.2 Codebook Example n Beams in
W.sub.1(n) W.sub.2 ([l.sub.1,n l.sub.2,n l.sub.3,n l.sub.4,n]) 0
[b.sub.0 b.sub.1 b.sub.2 b.sub.3] W.sub.2([0 0 0 0]) 1 [b.sub.2
b.sub.3 b.sub.4 b.sub.5] W.sub.2([1 1 0 0]) 2 [b.sub.4 b.sub.5
b.sub.6 b.sub.7] W.sub.2([1 1 0 0]) 3 [b.sub.6 b.sub.7 b.sub.8
b.sub.9] W.sub.2([1 1 0 0]) 4 [b.sub.8 b.sub.9 b.sub.10 b.sub.11]
W.sub.2([1 1 0 0]) 5 [b.sub.10 b.sub.11 b.sub.12 b.sub.13]
W.sub.2([1 1 0 0]) 6 [b.sub.12 b.sub.13 b.sub.14 b.sub.15]
W.sub.2([1 1 0 0]) 7 [b.sub.14 b.sub.15 b.sub.16 b.sub.17]
W.sub.2([1 1 0 0]) 8 [b.sub.16 b.sub.17 b.sub.18 b.sub.19]
W.sub.2([1 1 0 0]) 9 [b.sub.18 b.sub.19 b.sub.20 b.sub.21]
W.sub.2([1 1 0 0]) 10 [b.sub.20 b.sub.21 b.sub.22 b.sub.23]
W.sub.2([1 1 0 0]) 11 [b.sub.22 b.sub.23 b.sub.24 b.sub.25]
W.sub.2([1 1 0 0]) 12 [b.sub.24 b.sub.25 b.sub.26 b.sub.27]
W.sub.2([1 1 0 0]) 13 [b.sub.26 b.sub.27 b.sub.28 b.sub.29]
W.sub.2([1 1 0 0]) 14 [b.sub.28 b.sub.29 b.sub.30 b.sub.31]
W.sub.2([1 1 0 0]) 15 [b.sub.30 b.sub.31 b.sub.0 b.sub.1]
W.sub.2([1 1 1 1])
[0141] FIG. 15A illustrates an example of 2D co-polarized antenna
array 2000 according to embodiments of the present disclosure. The
embodiment of the 2D co-polarized antenna array 2000 shown in FIG.
15A is for illustration only. Other embodiments could be used
without departing from the scope of the present disclosure. FIG.
15A illustrates an example of 2D co-polarized antenna array 2000,
where MN virtual antenna ports, i.e., antenna ports (APs) 0 to
(MN-1) are placed on M.times.N 2D grid.
[0142] The 2D co-polarized antenna array 2000 includes a plurality
of antenna ports 2005. Each antenna port 2005 is mapped to a number
of physical antennas. In one implementation of 2D co-polarized
antenna array 2000, each antenna port 2005 is mapped to a single
antenna as depicted in FIG. 15B in which antenna port 2005 is
mapped to antenna 2055. Antenna 2055 is preceded by a power
amplifier block 2060, which is placed right after a carrier
modulation block 2065. Baseband transmit antenna input 2075 of size
N.sub.IFFT goes through an OFDM digital chain 2070, comprising
IFFT, P/S, add CP and DAC. The baseband transmit input signals for
the MN antenna ports can be constructed by pre-multiplying a
(MN.times.J) digital precoding matrix to J input signals, where
J.ltoreq.MN, as illustrated in FIG. 9.
[0143] FIG. 16 illustrates an example of 2D cross-polarized antenna
array 2100 according to embodiments of the present disclosure. The
embodiment of the 2D cross-polarized antenna array 2100 shown in
FIG. 16 is for illustration only. Other embodiments could be used
without departing from the scope of the present disclosure.
[0144] In the example shown in FIG. 16 the 2D cross-polarized
antenna array 2100 includes 2MN antenna ports, i.e., antenna ports
(APs) 0 to (2MN-1) are placed on M.times.N 2D grid, where each
point in the 2D grid has two antenna ports, one with +45.degree.
polarization and the other with -45.degree. polarization. In this
particular implementation of 2D co-polarized antenna array, each
antenna port comprises a single antenna. The baseband Tx input
signals for the 2MN antenna ports can be constructed by
pre-multiplying a (2MN.times.J) digital precoding matrix to J input
signals, where J.ltoreq.2MN.
[0145] FIG. 15A illustrates 2D co-pol arrays 2200 and 2250
according to embodiments of the present disclosure. The embodiments
of the 2D co-pol array 2200, 2250 shown in FIG. 17 are for
illustration only. Other embodiments could be used without
departing from the scope of the present disclosure.
[0146] The two 2D co-pol arrays 2200 and 2250 include the 2D
cross-polarized array shown in FIG. 16. The 2D co-polarized array
2250 includes the set of MN number of +45.degree. polarized
elements and the 2D co-polarized array 2200 includes the set of MN
number of -45.degree. polarized elements.
[0147] Definition (a unit norm complex number and DFT vector):
[0148] A complex number that can be mapped to a point at
2 .pi. L ##EQU00074##
radian phase angle on a unit circle is defined as
q L := - j 2 .pi. L . ##EQU00075##
[0149] An l-th DFT vector, b.sub.L,A(l) of length A, sampling
[0,2.pi.] with L levels are defined as in the following:
b.sub.L,A(l):=[1q.sub.L.sup.l, . . . , q.sub.L.sup.l(A-1)].sup.T:
l=0,1, . . . , L-1 (65).
[0150] Time and frequency domain baseband channel models:
[0151] For a 2D co-polarized uniform rectangular array (as
illustrated in FIGS. 20A and 20B and FIG. 17), a time-domain
baseband equivalent channel vector h(.tau.) (a MN.times.1 vector)
observed at a receive antenna at UE 116 is given by:
h(.SIGMA.)=.SIGMA..sub.l=0.sup.L-1.alpha..sub.la.sup.H(v.sub.l)a.sub.H(h-
.sub.l).delta.(.tau.-.tau..sub.l)e.sup.j2.pi.f.sup.c.sup..tau..sup.l(2)
(66).
where a(v.sub.l) is a M.times.1 elevation channel vector defined
as: a(v.sub.l)=[1e.sup.-jv.sup.l, . . . ,
e.sup.-j(M-1)v.sup.l].sup.T with
v i := 2 .pi. dv .lamda. c cos .theta. l , ##EQU00076##
and a(h.sub.l) is a N.times.1 azimuth channel vector defined as
a(h.sub.l)=[1e.sup.-jh.sup.l, . . . , e.sup.-j(N-1)h.sup.l].sup.T
with
h l := 2 .pi. d H .lamda. c sin .theta. l sin .PHI. l .
##EQU00077##
[0152] Accordingly, the frequency domain representation of the
model described in REF 2 is given by:
H(f)=.SIGMA..sub.l=0.sup.L-1.alpha..sub.la.sup.H(v.sub.l)a.sup.H(h.sub.l-
)e.sup.-j2.pi.(f-f.sup.c.sup.).tau..sup.l (67).
[0153] In the Line of Sight (LOS) case, the frequency domain
baseband equivalent model reduces to
H(f)=.alpha.a.sup.H(v)a.sup.H(h)e.sup.-j2.pi.(f-f.sup.c).tau..sup.0
(68).
where .alpha.:=.alpha..sub.0, v:=v and h:=h.sub.0. At a particular
frequency (or subcarrier), the spatial covariance matrix of H(f) is
given by:
R.sub.H=E[|.alpha.|.sup.2][a(v)a.sup.H(v)][a(h)a.sup.H(h)](3)
(69).
[0154] Definition: Rank-1 channel direction vector
[0155] Rank-1 channel direction matrix, or the principal
eigenvector .alpha. of R.sub.H is given by:
.alpha. = .beta. .alpha. ( v ) a ( h ) . where ( 70 ) a ( v ) = [ 1
- j v , , - j ( M - 1 ) v ] T with v := 2 .pi. d V .lamda. c cos
.theta. 0 ; and ( 71 ) a ( h ) = [ 1 - j h , , - j ( N - 1 ) h ] T
with h := 2 .pi. d H .lamda. c sin .theta. 0 sin .PHI. 0 . ( 72 )
##EQU00078##
with .theta..sub.0 being the elevation angle and .phi..sub.0 being
the azimuth angle, and .beta. is a normalized factor.
[0156] Certain embodiments include, based on the structure of the
spatial covariance matrix R.sub.H, a codebook based on the
Kronecker product of two DFT vectors. It is intended that the two
DFT vectors respectively quantize conjugates of the elevation and
the azimuth channel vectors a(v)=[1e.sup.-jv, . . . ,
e.sup.-j(M-1)v].sup.T and a(h)=[1e.sup.-jh, . . . ,
e.sup.-j(N-1)h].sup.T that construct the rank-1 channel direction
vector .alpha.=.beta.a(v)a(h).
[0157] NodeB 102 antenna spacing configuration for configuring a
codebook
[0158] In order to quantize a(v) and a(h), either the elevation
and/or azimuth angles .theta..sub.0 and .phi..sub.0, or their
exponents v and h, can be quantized. An advantage of the exponent
quantization is that UE 116 can form horizontal and vertical
pre-coding vectors from the codebook (since they are DFT vectors)
without knowing the horizontal and vertical antenna spacing
(d.sub.V and d.sub.H) at NodeB 102. Alternatively, for angle
quantization, antenna spacing needs to be configured to UE 116 so
that UE 116 can construct the horizontal and vertical pre-coding
vectors from the codebook. This configuration can be sent to UE 116
via higher layer signaling such as RRC. The horizontal and vertical
antenna spacings (d.sub.V and d.sub.H) can take different values
based on the configuration. In one example, NodeB 102 can configure
at least one of d.sub.H/.lamda..sub.c and d.sub.V.lamda..sub.c
values, each chosen from 4 candidate values as shown below.
TABLE-US-00004 TABLE 4 Horizontal antenna spacing {0.5, 1.0, 2.0,
4.0} (d.sub.H/.lamda..sub.c) Vertical antenna spacing
(d.sub.V/.lamda..sub.c) {0.5, 1.0, 2.0, 4.0}
[0159] The antenna spacing information can be either
cell-specifically or UE-specifically configured.
[0160] Unless otherwise specified, a closely spaced antenna
configuration is assumed throughout the remainder of the
disclosure, where d.sub.H=d.sub.V=0.5.lamda..sub.c are considered,
v:=.pi. cos .theta..sub.0 and h:=.pi. sin .theta..sub.0 sin
.phi..sub.0. Having these observations, a product codebook with
independently quantizing two channel vectors a(v) and a(h) can be
devised, as in the following.
[0161] A rank-1 codebook (W) for 2D co-polarized array with
uniformly quantizing the exponents h=.pi. sin .theta..sub.0 sin
.phi..sub.0 and v=.pi. cos .theta..sub.0. A codeword w is indexed
by two non-negative integers, n and m, and represented by a
Kronecker product:
w(n,m)=.beta.w.sub.v(m)w.sub.h(n),w.sub.v(m).epsilon.W.sub.v,w.sub.h(n).-
epsilon.W.sub.h
[0162] In one method, W.sub.v is defined as a set of DFT vectors
uniformly sampling [0,2.pi.] with P.sub.v quantization levels, and
W.sub.h is defined as a set of DFT vectors uniformly sampling
[0,2.pi.] with P.sub.v, and Q.sub.h
quantization levels:
W.sub.v={b.sub.P.sub.v.sub.,M(m):m=0, . . . , P.sub.v-1}; and
W.sub.h={b.sub.Q.sub.h.sub.,N(n):n=0, . . . , Q.sub.h-1}.
[0163] In this method, e.sup.-jv and e.sup.-jh of the rank-1
channel direction vectors are respectively quantized by
q P v = - j 2 .pi. n P v ##EQU00079## and ##EQU00079.2## q Q h = -
j 2 .pi. n Q h . ##EQU00079.3##
In one example P.sub.v=4 and Q.sub.h=16 and M=N=4; then
W v = { w v ( n ) = [ 1 j 2 .pi. n 4 j 2 .pi. 2 n 4 j 2 .pi. 3 n 4
] T : n = 0 , 1 , 2 , 3 } ; and ( 73 ) W h = { w h ( n ) = [ 1 j 2
.pi. n 16 j 2 .pi. 2 n 16 j 2 .pi. 3 n 16 ] T : n = 0 , 1 , 2 , 15
} . ( 74 ) . ##EQU00080##
[0164] An example UE implementation on PMI selection: in certain
embodiments, UE 116 selects v-PMI and h-PMI by maximizing the
cosine of the hyper angle formed by the two vectors H(f) and
w.sub.1, as in the following:
= arg max w 1 .di-elect cons. W v W h = arg max m = 0 , 1 , , P v -
1 , n = 0 , 1 , , Q h - 1 a H ( v ) w v ( m ) 2 a ( v ) 2 w v ( m )
2 a H ( h ) w h ( n ) 2 a ( h ) 2 w h ( n ) 2 . ( 75 ) .
##EQU00081##
CSI feedback: in certain embodiments, NodeB 102 configures UE 116
to report two PMI indices, a vertical PMI (v-PMI), an index of a
codeword in v-codebook and a horizontal PMI (h-PMI), an index of a
codeword in h-codebook. Upon receiving the two PMI indices, NodeB
102 constructs a rank-1 channel by taking a Kronecker product of
the two precoding vectors indicated by the two PMI indices. For a
set of antenna ports L, a v-PMI value of m.epsilon.{0, 1, . . . ,
P.sub.v-1} and a h-PMI value of n.epsilon.{0, 1, . . . , Q.sub.h-1}
corresponds to the codebook indices m and n given in Table.
TABLE-US-00005 TABLE 5 Codebook for 1-layer CSI reporting using a
set of V and H antenna ports h-PMI v-PMI n m 0 1 . . . n . . .
Q.sub.m - 1 {0, 1, . . . , P.sub.v - 1} W.sub.m,0 W.sub.m,1 . . .
W.sub.m,n . . . W.sub.m,Qm-1 w(n, m) = .beta.w.sub.v(m)
w.sub.h(n)
[0165] Dependency of azimuth angle distribution on elevation angle.
To understand the dependency of two PMIs, an alternative expression
for the parameter h.sub.mn is derived as in FIG. 18. Let
.alpha..sub.0 be the angle between the direction of departure and
the positive y-axis direction. As can be seen from FIG. 18, the
following geometric relationships between .theta..sub.0,
.phi..sub.0, and .alpha..sub.0 can be obtained:
sin .theta. 0 = x 0 2 + y 0 2 x 0 2 + y 0 2 + z 0 2 , sin .PHI. 0 =
y 0 x 0 2 + y 0 2 , and cos .alpha. 0 = y 0 x 0 2 + y 0 2 + z 0 2 .
( 76 ) ##EQU00082##
Thus, cos .alpha..sub.0=sin .theta..sub.0 sin .phi..sub.0. The
parameter v.sub.0 can be rewritten as
v 0 = 2 .pi. d V .lamda. c cos .alpha. 0 . ##EQU00083##
[0166] Alternative definition: Rank-1 channel direction vector:
[0167] In its alternative definition according to the observation
here, rank-1 channel direction matrix, or the principal eigenvector
.alpha. of R.sub.H is given by:
.alpha. = .beta. a ( v ) a ( h ) . where ( 77 ) a ( v ) = [ 1 - j v
, , - j ( M - 1 ) v ] T with v := 2 .pi. d V .lamda. c cos .theta.
0 ( 78 ) a ( h ) = [ 1 - j h , , - j ( N - 1 ) h ] T ; with h := 2
.pi. d H .lamda. c cos .alpha. 0 . ( 79 ) ##EQU00084##
with .theta..sub.0 being the elevation angle and .alpha..sub.0
being the angle between the direction of departure and the positive
y-axis direction, and .beta. is a normalizing factor.
[0168] FIG. 19 illustrates the relationship between angles
.alpha..sub.0 and .theta..sub.0 according to embodiments of the
present disclosure. The embodiment of the relationship shown in
FIG. 19 is for illustration only. Other embodiments could be used
without departing from the scope of the present disclosure.
[0169] Referring to FIG. 19, the relation between the parameters
.theta..sub.0 and .alpha..sub.0 satisfy the following
constraints:
For 0 .ltoreq. .theta. 0 .ltoreq. .pi. 2 , .pi. 2 - .theta. 0
.ltoreq. .alpha. 0 .ltoreq. .pi. 2 + .theta. 0 ( 80 ) For .pi. 2
.ltoreq. .theta. 0 .ltoreq. .pi. , .theta. 0 - .pi. 2 .ltoreq.
.alpha. 0 .ltoreq. 3 .pi. 2 - .theta. 0 ( 81 ) ##EQU00085##
[0170] FIG. 20 illustrates the feasible region of .theta. and
.alpha. according to embodiments of the present disclosure. The
embodiment of the feasible region shown in FIG. 20 is for
illustration only. Other embodiments could be used without
departing from the scope of the present disclosure.
[0171] In the example shown in FIG. 20, the parameters .theta. and
.alpha. are constrained within the shaded (red) region. Note that
if
.theta. = .pi. 2 , ##EQU00086##
then UE 116 is in the XY plane, hence .alpha. can take any value in
[0,.pi.]. Alternatively, for .theta.=0 or .pi., .alpha. can take
only one value
.alpha. = .pi. 2 . ##EQU00087##
In general, the length of range of possible values of .alpha.
decreases for .theta. values away from
.pi. 2 . ##EQU00088##
Also, this behavior is symmetric about
.theta. = .pi. 2 . ##EQU00089##
FIG. 20 further illustrates the potential benefits of exploiting
the dependency of two PMIs: either the feedback overhead can be
reduced or the throughput performance can be improved.
[0172] FIG. 21 illustrates the dependency of the exponents v and h
when .phi..sub.0.epsilon.[-.pi.,.pi.t] according to embodiments of
the present disclosure. The embodiment of the dependency shown in
FIG. 21 is for illustration only. Other embodiments could be used
without departing from the scope of the present disclosure.
[0173] According to its alternative definition, rank-1 channel
direction vectors are obtained with the cosine values of .theta.
and .alpha., i.e. .pi. cos .theta. and .pi. cos .alpha.. In FIG.
21, these cosine values are plotted where each point represents a
possible pair of (v, h) exponents. The white rectangle consists of
all possible the (v, h) pairs without considering the constraints
in equations 80 and 81 while the gray oval consists of the (v, h)
pairs that takes the constraints in equations 80 and 81 into
account. Clearly, the area of the oval is smaller than that of the
rectangle. It implies that there exists a potential saving if the
dependency of the v-PMI and the h-PMI is exploited.
[0174] Codebook 2: A rank-1 codebook (W) for 2D co-polarized array
with exploiting the dependency between angles .theta. and
.alpha.:
Example 1
[0175] A codeword w is indexed by two non-negative integers, m and
n, and represented by a Kronecker product:
w(m,n)=.beta.w.sub.v(m)w.sub.h(m,n),w.sub.v(m).epsilon.W.sub.v,w.sub.h(m-
,n).epsilon.W.sub.h(m)
W.sub.v and W.sub.h are defined as two sets of DFT precoding
vectors:
W.sub.v={b.sub.P.sub.v.sub.,M(m): m=0, . . . , P.sub.v-1}; and
W.sub.h={b.sub.Q.sub.h.sub.,N(M,n): n.epsilon.S.sub.m}.
[0176] In this method, e.sup.-jv and e.sup.-jh of the rank-1
channel direction vectors are respectively quantized by
q P v = - j v m = - j 2 m .pi. P v ##EQU00090##
and q.sub.Q.sub.h=e.sup.-jh.sup.mn, where h.sub.mn is obtained by
uniformly quantizing the range of possible values of .alpha. into
Q.sub.h levels based on the v-PMI index m. Furthermore, codeword
w.sub.v is determined based upon a single index m. However codeword
w.sub.h is determined based upon index m as well as n, motivated by
the dependency of .alpha. (or exponent h) on .theta. (exponent
v).
[0177] FIG. 22 illustrates a high-rise scenario according to
embodiments of the present disclosure. The embodiment of the
high-rise scenario shown in FIG. 22 is for illustration only. Other
embodiments could be used without departing from the scope of the
present disclosure.
[0178] PMI construction example 1: 2-bit uniform sampling of
[0,2.pi.] for v.
[0179] In one example construction of codebook 2, P.sub.v=4 number
of states are used for uniformly quantizing v=.pi. cos
.theta.+.pi..epsilon.[0,2.pi.] utilizing
q 4 = - j 2 .pi. 4 ##EQU00091##
and corresponding DFT vector codebook W.sub.v={b.sub.4,M(m): m=0,
1, 2, 3}. Then, a quantization range of h=.pi. cos .alpha.+.pi. is
determined dependent upon the selected value of v and .theta.,
according to the feasibility region in FIG. 22. Table 6 shows the
resulting PMI table, where Q.sub.h=16, for example. When this
codebook is configured, a UE selects w(m, n) and corresponding
indices m and n maximizing the channel quality, and feeds back the
indices in and n.
TABLE-US-00006 TABLE 6 PMI construction example 1 v-PMI index m v m
= 2 m .pi. 4 ##EQU00092## .theta..sub.m Quantization range of
.alpha. Quantization range of h h-PMI index n h.sub.m,n 0 0 0 [
.pi. 2 , .pi. 2 ] ##EQU00093## [0, 0] n .epsilon. {0, 1, . . . ,
Q.sub.h - 1} 0 1 .pi. 2 ##EQU00094## 2 .pi. 3 ##EQU00095## [ .pi. 6
, 5 .pi. 6 ] ##EQU00096## [0, {square root over (3)}.pi.] n
.epsilon. {0, 1, . . . , Q.sub.h - 1} .pi. 6 + 3 .pi. Q h
##EQU00097## 2 .pi. .pi. 2 ##EQU00098## [0, .pi.] [0, 2.pi.] n
.epsilon. {0, 1, . . . , Q.sub.h - 1} 2 .pi. n Q h ##EQU00099## 3 3
.pi. 2 ##EQU00100## .pi. 3 ##EQU00101## [ .pi. 6 , 5 .pi. 6 ]
##EQU00102## [0, {square root over (3)}.pi.] n .epsilon. {0, 1, . .
. , Q.sub.h - 1} .pi. 6 + 3 .pi. Q h ##EQU00103##
[0180] Codebook 2: A rank-1 codebook (W) for 2D co-polarized array
with exploiting the dependency between angles .theta. and .alpha.:
example 2. A codeword w is indexed by two non-negative integers, m
and n, and represented by a Kronecker product:
w(m,n)=.beta.w.sub.v(m)w.sub.h(m,n),w.sub.v(m).epsilon.W.sub.v,w.sub.h(m-
,n).epsilon.W.sub.h(m)
W.sub.v, and W.sub.h are defined as two sets of DFT precoding
vectors:
W.sub.v={w.sub.v(m)=[1e.sup.-jv.sup.m, . . . ,
e.sup.-j(M-1)v.sup.m].sup.T: m=0, . . . , P.sub.v-1}; and
W.sub.h={b.sub.Q.sub.h.sub.,N(M,n):n.epsilon.S.sub.m} (82).
Where v.sub.m:=.pi.+.pi. cos .theta..sub.m.
[0181] In this method, e.sup.-jv and e.sup.-jh of the rank-1
channel direction vectors are respectively quantized by v.sub.m:
=.pi.+.pi. cos .theta..sub.m and h.sub.mn which is obtained by
uniformly quantizing the range of possible values of .alpha. into
Q.sub.h levels based on the v-PMI index m. Furthermore, codeword
w.sub.v is determined based upon a single index m. However codeword
w.sub.h is determined based upon index m as well as n, motivated by
the dependency of .alpha. (or exponent h) on .theta. (or exponent
v).
[0182] PMI construction example 2: 4-bit uniform sampling of
[0,.pi.t] for .theta..sub.m, fixed number of bits for h-channel
quantization.
[0183] In one example construction of codebook 2, P.sub.v=16 number
of states are used for uniformly quantizing
.theta..epsilon.[0,.pi.] with
.theta. m = m .pi. 16 . ##EQU00104##
Then the resulting
v m = .pi. + .pi. cos m .pi. 16 . ##EQU00105##
In addition, for h-channel quantization, Q.sub.h=16 is the number
of states as shown in Table 12.
TABLE-US-00007 TABLE 7 PMI construction example 2 ( with v m = .pi.
+ .pi. cos .theta. m , and .theta. m = m .pi. 16 ) ##EQU00106##
v-PMI index m .theta. m = m .pi. 16 ##EQU00107## Quantization range
of .alpha. Quantization range of h h-PMI index n h.sub.m,n 0 0 [ 8
.pi. 16 , 8 .pi. 16 ] ##EQU00108## [0, 0] n .epsilon. {0, 1, . . .
, Q.sub.h - 1} 0 1 .pi. 16 ##EQU00109## [ 7 .pi. 16 , 9 .pi. 16 ]
##EQU00110## [ 0 , 2 .pi. cos 7 .pi. 16 ] ##EQU00111## n .epsilon.
{0, 1, . . . , Q.sub.h - 1} 7 .pi. 16 + ( 2 .pi. n cos 7 .pi. 16 )
/ Q h ##EQU00112## 2 2 .pi. 16 ##EQU00113## [ 6 .pi. 16 , 10 .pi.
16 ] ##EQU00114## [ 0 , 2 .pi. cos 6 .pi. 16 ] ##EQU00115## n
.epsilon. {0, 1, . . . , Q.sub.h - 1} 6 .pi. 16 + ( 2 .pi. n cos 6
.pi. 16 ) / Q h ##EQU00116## 3 3 .pi. 16 ##EQU00117## [ 5 .pi. 16 ,
11 .pi. 16 ] ##EQU00118## [ 0 , 2 .pi. cos 5 .pi. 16 ] ##EQU00119##
n .epsilon. {0, 1, . . . , Q.sub.h - 1} 5 .pi. 16 + ( 2 .pi. n cos
5 .pi. 16 ) / Q h ##EQU00120## 4 4 .pi. 16 ##EQU00121## [ 4 .pi. 16
, 12 .pi. 16 ] ##EQU00122## [ 0 , 2 .pi. cos 4 .pi. 16 ]
##EQU00123## n .epsilon. {0, 1, . . . , Q.sub.h - 1} 4 .pi. 16 + (
2 .pi. n cos 4 .pi. 16 ) / Q h ##EQU00124## 5 5 .pi. 16
##EQU00125## [ 3 .pi. 16 , 13 .pi. 16 ] ##EQU00126## [ 0 , 2 .pi.
cos 3 .pi. 16 ] ##EQU00127## n .epsilon. {0, 1, . . . , Q.sub.h -
1} 3 .pi. 16 + ( 2 .pi. n cos 3 .pi. 16 ) / Q h ##EQU00128## 6 6
.pi. 16 ##EQU00129## [ 2 .pi. 16 , 14 .pi. 16 ] ##EQU00130## [ 0 ,
2 .pi. cos 2 .pi. 16 ] ##EQU00131## n .epsilon. {0, 1, . . . ,
Q.sub.h - 1} 2 .pi. 16 + ( 2 .pi. n cos 2 .pi. 16 ) / Q h
##EQU00132## 7 7 .pi. 16 ##EQU00133## [ .pi. 16 , 15 .pi. 16 ]
##EQU00134## [ 0 , 2 .pi. cos .pi. 16 ] ##EQU00135## n .epsilon.
{0, 1, . . . , Q.sub.h - 1} .pi. 16 + ( 2 .pi. n cos .pi. 16 ) / Q
h ##EQU00136## 8 8 .pi. 16 ##EQU00137## [0, .pi.] [0, 2.pi.] n
.epsilon. {0, 1, . . . , Q.sub.h - 1} (2.pi.n)/Q.sub.h 9 9 .pi. 16
##EQU00138## [ .pi. 16 , 15 .pi. 16 ] ##EQU00139## [ 0 , 2 .pi. cos
.pi. 16 ] ##EQU00140## n .epsilon. {0, 1, . . . , Q.sub.h - 1} .pi.
16 + ( 2 .pi. n cos .pi. 16 ) / Q h ##EQU00141## 10 10 .pi. 16
##EQU00142## [ 2 .pi. 16 , 14 .pi. 16 ] ##EQU00143## [ 0 , 2 .pi.
cos 2 .pi. 16 ] ##EQU00144## n .epsilon. {0, 1, . . . , Q.sub.h -
1} 2 .pi. 16 + ( 2 .pi. n cos 2 .pi. 16 ) / Q h ##EQU00145## 11 11
.pi. 16 ##EQU00146## [ 3 .pi. 16 , 13 .pi. 16 ] ##EQU00147## [ 0 ,
2 .pi. cos 3 .pi. 16 ] ##EQU00148## n .epsilon. {0, 1, . . . ,
Q.sub.h - 1} 3 .pi. 16 + ( 2 .pi. n cos 3 .pi. 16 ) / Q h
##EQU00149## 12 12 .pi. 16 ##EQU00150## [ 4 .pi. 16 , 12 .pi. 16 ]
##EQU00151## [ 0 , 2 .pi. cos 4 .pi. 16 ] ##EQU00152## n .epsilon.
{0, 1, . . . , Q.sub.h - 1} 4 .pi. 16 + ( 2 .pi. n cos 4 .pi. 16 )
/ Q h ##EQU00153## 13 13 .pi. 16 ##EQU00154## [ 5 .pi. 16 , 11 .pi.
16 ] ##EQU00155## [ 0 , 2 .pi. cos 5 .pi. 16 ] ##EQU00156## n
.epsilon. {0, 1, . . . , Q.sub.h - 1} 5 .pi. 16 + ( 2 .pi. n cos 5
.pi. 16 ) / Q h ##EQU00157## 14 14 .pi. 16 ##EQU00158## [ 6 .pi. 16
, 10 .pi. 16 ] ##EQU00159## [ 0 , 2 .pi. cos 6 .pi. 16 ]
##EQU00160## n .epsilon. {0, 1, . . . , Q.sub.h - 1} 6 .pi. 16 + (
2 .pi. n cos 6 .pi. 16 ) / Q h ##EQU00161## 15 15 .pi. 16
##EQU00162## [ 7 .pi. 16 , 9 .pi. 16 ] ##EQU00163## [ 0 , 2 .pi.
cos 7 .pi. 16 ] ##EQU00164## n .epsilon. {0, 1, . . . , Q.sub.h -
1} 7 .pi. 16 + ( 2 .pi. n cos 7 .pi. 16 ) / Q h ##EQU00165##
[0184] PMI construction example 3: 4-bit uniform sampling of
[0,.pi.] for .theta..sub.m, the number of bits for h-channel
quantization determined as a function of .theta..sub.m: In one
example construction of codebook 2 shown in Table 8, P.sub.v=16
number of states are used for uniformly quantizing
.theta..epsilon.[0,.pi.] with
.theta. m = m .pi. 16 . ##EQU00166##
Then the resulting
v m = .pi. + .pi.cos m .pi. 16 . ##EQU00167##
Alternatively, the number of states for h-channel quantization,
Q.sub.h is determined as a function of m. For example, when m=0,
only one value is used for h-PMI (Q.sub.h=1), and when m=2, 3
values are used for h-PMI (Q.sub.h=3). The PMI indices (m,n) are
fed back by UE 116 to NodeB 102 by means of a composite PMI i as
shown in Table 14.
TABLE-US-00008 TABLE 8 Codebook for 1-layer CSI reporting using a
set of antenna ports L v-PMI 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
(m) Q.sub.h 1 1 3 3 5 5 7 7 9 7 7 5 5 3 3 1 h-PMI 4 4 3 3 2 2 1 1 0
1 1 2 2 3 3 4 (n) 4 4 3 3 2 2 1 2 2 3 3 4 4 5 5 4 4 3 3 2 3 3 4 4 5
5 5 5 4 4 3 4 4 5 5 6 6 5 5 4 5 5 6 6 6 6 5 6 6 7 7 6 7 7 7 8
TABLE-US-00009 TABLE 9 Composite PMI for 1-layer CSI Reporting
Composite (H, V) PMI PMI index i index 0 (0, 4) 1 (1, 4) 2 (2, 3) 3
(2, 4) 4 (2, 5) 5 (3, 3) 6 (3, 4) 7 (3, 5) 8 (4, 2) 9 (4, 3) 10 (4,
4) 11 (4, 5) 12 (4, 6) 13 (5, 2) 14 (5, 3) 15 (5, 4) 16 (5, 5) 17
(5, 6) 18 (6, 1) 19 (6, 2) 20 (6, 3) 21 (6, 4) 22 (6, 5) 23 (6, 6)
24 (6, 7) 25 (7, 1) 26 (7, 2) 27 (7, 3) 28 (7, 4) 29 (7, 5) 30 (7,
6) 31 (7, 7) 32 (8, 0) 33 (8, 1) 34 (8, 2) 35 (8, 3) 36 (8, 4) 37
(8, 5) 38 (8, 6) 39 (8, 7) 40 (9, 1) 41 (9, 2) 42 (9, 3) 43 (9, 4)
44 (9, 5) 45 (9, 6) 46 (9, 7) 47 (10, 1) 48 (10, 2) 49 (10, 3) 50
(10, 4) 51 (10, 5) 52 (10, 6) 53 (10, 7) 54 (11, 2) 55 (11, 3) 56
(11, 4) 57 (11, 5) 58 (11, 6) 59 (12, 2) 60 (12, 3) 61 (12, 4) 62
(12, 5) 63 (12, 6) 64 (13, 3) 65 (13, 4) 66 (13, 5) 67 (14, 3) 68
(14, 4) 69 (14, 5) 70 (15, 4)
[0185] Codebook subsampling of PMI example 3:
[0186] In LTE/LTE-Advanced downlink, UE 116 reports CQI, PMI, or RI
to NodeB 102 via a feedback channel. There are two types of
feedback channels: PUCCH and PUSCH. Periodic CSI is transmitted via
PUCCH in a semi-statically configured manner while aperiodic CSI is
transmitted via PUSCH in a dynamic manner. Compared with PUSCH,
PUCCH has a more stringent requirement on the payload size. For
example, in Release 12, period CSI is transmitted using PUCCH
format 2/2a/2b with a payload size up to 11 bits. In Release 12,
codebook subsampling is used to meet the PUCCH payload size
requirement. Another method to reduce overhead and complexity in
LTE is PMI codebook subset restriction (CSR). In CSR, any codeword
can be disabled via CSR bitmap in radio resource control (RRC)
signaling. UE 116 only needs to search the restricted codebook
subset for PMI reporting. However, a large codebook size leads to a
large overhead related to the transmission of a long bitmap. Seeing
these issues, here we propose to use the PMI codebook 3 shown in
Table 13 and Table 14 for PUSCH feedback, and to use a subsampled
version of the PMI codebook 3 for PUCCH feedback.
[0187] In one example, the pairs of (m,n) of the PMI codebook 3 are
uniformly subsampled in two dimension as shown in Table. Out of
this subsampling, the total number of states to be fed back is
reduced from 70 to 25, in which 25 states can fit in 5 bits. The
composite PMI table to map composite PMI index i to (m,n) can be
constructed for Table 15, as done for PMI codebook 3.
TABLE-US-00010 TABLE 10 Codebook for 1-layer CSI reporting using a
set of antenna ports L v-PMI 0 2 4 6 8 10 12 14 15 (m) Q.sub.h 1 2
3 4 5 4 3 2 1 h-PMI 4 3 2 1 0 1 2 3 4 (n) 5 4 3 2 3 4 5 6 5 4 5 6 7
6 7 8
[0188] PMI construction example 4: non-uniform quantization of
.theta.. In PMI construction example 4 in Table 16, .theta. is
non-uniformly quantized with values in a set of {.pi./4, .pi./2,
5.pi./8, 3.pi./4}. In the urban macro/micro scenarios, the
distribution of elevation angle .theta. is concentrated around
.pi./2, and there are richer distributions for .theta.>.pi./2
than .theta.<.pi./2. To reflect this distribution, certain
embodiments allocate more states for .theta.>.pi./2 than
.theta.<.pi./2. In this example, 2 states {5.pi./8, 3.pi./4} are
allocated to quantize .theta.>.pi./2, and a single state
{.pi./4} is allocated to quantize .theta.<.pi./2.
TABLE-US-00011 TABLE 11 PMI construction example 4 v-PMI
Quantization index range m .theta..sub.m of .alpha. h-PMI index n
h.sub.m,n 0 .pi. 4 ##EQU00168## [ .pi. 4 , 3 .pi. 4 ] ##EQU00169##
n .epsilon. {0, 1, . . . , Q.sub.h - 1} .pi. 4 + 2 .pi. n Q h
##EQU00170## .pi. 2 ##EQU00171## [0, .pi.] n .epsilon. {0, 1, . . .
, Q.sub.h - 1} 2 .pi. n Q h ##EQU00172## 5 8 .pi. ##EQU00173## [
.pi. 8 , 7 .pi. 8 ] ##EQU00174## n .epsilon. {0, 1, . . . , Q.sub.h
- 1} .pi. 8 + 2 + 2 .pi. n Q h ##EQU00175## 3 .pi. 4 ##EQU00176## [
.pi. 4 , 3 .pi. 4 ] ##EQU00177## n .epsilon. {0, 1, . . . , Q.sub.h
- 1} .pi. 4 + 2 .pi. n Q h ##EQU00178##
[0189] Configurable rank-1 codebook:
[0190] Referring to FIG. 22, a high-rise scenario is currently
considered in the 3GPP RAN 1 meeting. The elevation angle can be as
small as 30 degrees. This embodiment presents a codebook W that has
two sub-codebooks. One sub-codebook can be configured for those UEs
located in low-rise buildings while the other sub-codebook can be
configured for those UEs located in high-rise buildings. A
sub-codebook can be configured by NodeB 102 via RRC signaling,
either UE specifically or cell specifically. The two sub-codebook
may have different resolutions. In one example, low-rise and
high-rise are referred to as the scenarios where the elevation
angles are restricted to
[ 5 .pi. 12 , 7 .pi. 12 ] and [ .pi. 6 , 5 .pi. 12 ] ,
##EQU00179##
respectively. The low-rise sub-codebook W.sub.LR has coarse
codebook resolution on both the vertical and horizontal domains as
compared with the high-rise sub-codebook W.sub.HR. The low-rise
sub-codebook W.sub.LR with P.sub..theta.=Q.sub..alpha.=16 is given
by:
W LR = { .beta. a ( u m ) a ( s n ) : 0 .ltoreq. 5 .pi. 12 + m .pi.
96 .ltoreq. .pi. 2 , .pi. 12 - m .pi. 96 .ltoreq. n .pi. 8 .ltoreq.
.pi. 12 + m .pi. 96 } { .beta. a ( u m ) ( s n ) : .pi. 2 .ltoreq.
5 .pi. 12 + m .pi. 96 .ltoreq. .pi. , .pi. 12 + m .pi. 96 .ltoreq.
n .pi. 8 .ltoreq. 25 .pi. 12 - m .pi. 96 } ##EQU00180## where
##EQU00180.2## v m = .pi. cos ( 5 .pi. 12 + m .pi. 96 )
##EQU00180.3## and ##EQU00180.4## s n = .pi. cos ( n .pi. 8 ) ,
##EQU00180.5##
while the high-rise sub-codebook W.sub.HR with
P.sub..theta.=Q.sub..alpha.=32 is given by
W HR = { .beta. a ( u m ) a ( s n ) : 0 .ltoreq. .pi. 6 + m .pi.
128 .ltoreq. .pi. 2 , .pi. 3 - m .pi. 128 .ltoreq. n .pi. 16
.ltoreq. 2 .pi. 3 + m .pi. 128 } { .beta. a ( u m ) a ( s n ) :
.pi. 2 .ltoreq. .pi. 6 + m .pi. 128 .ltoreq. .pi. , 2 .pi. 3 + m
.pi. 128 .ltoreq. n .pi. 16 .ltoreq. 7 .pi. 3 - m .pi. 96 } where v
m = .pi. cos ( .pi. 6 + m .pi. 128 ) and s n = .pi. cos ( n .pi. 16
) . ( 83 ) . ##EQU00181##
TABLE-US-00012 TABLE 12 An Example of RRC Configured Two
Sub-codebooks Scenarios Codebooks Low - rise : .theta. .di-elect
cons. [ 5 .pi. 12 , 7 .pi. 12 ] ##EQU00182## Sub-codebook W.sub.LR
High - rise : .theta. .di-elect cons. ( .pi. 6 , 5 .pi. 12 ]
##EQU00183## Sub-codebook W.sub.HR
[0191] FIG. 23 illustrates quantization of cosine values of the
elevation angles according to embodiments of the present
disclosure. The embodiment of the quantization of cosine values of
the elevation angles shown in FIG. 23 is for illustration only.
Other embodiments could be used without departing from the scope of
the disclosure.
[0192] FIG. 23 plots .theta. versus cos .theta. and the `x` stands
for the coordinate (cos .theta..sub.i,.theta..sub.i) satisfying cos
.theta..sub.i-cos .theta..sub.i+1=c for a constant c and i=0, . . .
, 31. As evident in FIG. 23, the elevation angles are not quantized
uniformly. For applications such as multi-user MIMO (MU-MIMO), this
non-uniform quantization may degrade performance. For example, when
two users happen to have elevation angles in [0, 20] degrees, the
two users' channels can be quantized with the same PMI index. In
this case, NodeB 102 cannot MU-MIMO multiplex these two UEs, such
as UE 115 and UE 116, together in the same time-frequency resource.
The NodeB 102 is unable to MU-MIMO multiplex UE 115 and UE 116, not
because the channel does not support the MU-MIND multiplexing, but
due to a lack of information.
[0193] To improve the performance, certain embodiments construct a
rank-1 co-pol codebook by quantizing azimuth and elevation angles
uniformly.
[0194] Codebook Construction
[0195] Certain embodiments quantize the elevation angle .theta.
into P.sub..theta. equal quantization levels and quantize the
azimuth angle .phi. into Q.sub..phi. equal quantization levels. The
elevation angle can be .theta..epsilon.[E.sub.l,E.sub.u) and the
azimuth angle can be .phi..epsilon.[A.sub.l, A.sub.u), where
E.sub.l and E.sub.u are the lower- and upper-limits of the
elevation angle, respectively, and satisfy
0.ltoreq.E.sub.l<E.sub.u.ltoreq..pi., and A.sub.l and A.sub.u
are the lower- and upper-limits of the azimuth angle, respectively
and satisfy 0.ltoreq.A.sub.l<A.sub.u.ltoreq.2.pi.. The feasible
range of the elevation and azimuth angles may depend on factors
such as UE distributions and antenna configurations.
[0196] The elevation precoding vector is given by:
f ( v m ) := [ 1 - j v m , , - j ( M - 1 ) v m ] T with v m = 2
.pi. d V .lamda. c cos ( E l + m ( E u - E l ) P .theta. ) . ( 84 )
##EQU00184##
for m=0, 1, . . . , P.sub..theta.-1, and the azimuth precoding
vector is given by:
f ( h m , m ) := [ 1 - j v n , m , , - j ( N - 1 ) h m , n ] T with
h m , n = 2 .pi. d H .lamda. c sin ( E l + m ( E u - E l ) P
.theta. ) sin ( A l + n ( A u - A l ) Q .PHI. ) . ( 85 )
##EQU00185##
for m=0, 1, . . . , P.sub..theta.-1 and n=0, 1, . . . ,
Q.sub..phi.-1. The overall rank-1 W codebook can be expressed
as
W={W.sub.m,n=.beta.f(v.sub.m)f(h.sub.m,n):m=0, 1, . . . ,
P.sub..theta.-1, n=0, 1, . . . , Q.sub..phi.-1} (86).
.beta. is a normalization factor. The number of codewords in this
codebook is P.sub..theta.Q.sub..phi..
[0197] Codebook Construction Example 1: Consider a 8V.times.8H 2D
URA (as shown in FIG. 15A or FIG. 17). Assume that the feasible
ranges of elevation and azimuth angles are .theta..epsilon.[0,
.pi.] and .phi..epsilon.[0,2.pi.]. The values of parameters are
given by P.sub..theta.=4, Q.sub..phi.=16, which leads to a 2-bit
vertical codebook and a 4-bit horizontal codebook. With plugging
d.sub.H=d.sub.V=0.5.lamda..sub.c, the parameters
v m = .pi. cos m .pi. 4 ##EQU00186## and ##EQU00186.2## h m , n =
.pi. sin m .pi. 4 sin 2 mn .pi. 16 ##EQU00186.3##
with m=0, 1, . . . , 3 and n=0, . . . , 15. A codeword W.sub.m,n is
given by:
W m , n = 1 16 f ( v m ) f ( h m , n ) = 1 16 [ 1 q m , n q m , n
15 p m p m q m , n 15 p m 15 q m , n 15 ] where p m = j.pi. cos ( m
.pi. 16 ) and q m , n = j.pi. sin ( m .pi. 16 ) sin ( 2 mn .pi. 16
) . ( 87 ) ##EQU00187##
[0198] PMI Feedback
[0199] For co-pol 2D antenna array, a v-PMI value of m.epsilon.{0,
1, . . . , P.sub..theta.-1} and a h-PMI value of n.epsilon.{0, 1, .
. . , Q.sub..phi.-1} corresponds to the codebook indices m and n
given in Table 13. When configured with this PMI table, UE 116
needs to feeds back h-PMI n and v-PMI m to the NodeB 102.
TABLE-US-00013 TABLE 13 Codebook for 1-layer for co-pol 2D antenna
array h-PMI v-PMI n m 0 1 . . . n . . . Q.sub..phi. - 1 m W.sub.m,0
W.sub.m,1 . . . W.sub.m,n . . . W.sub.m,Q.phi.-1 .di-elect cons.
{0, 1, . . . , P.sub..theta. - 1} W.sub.m,n = .beta.f(v.sub.m)
f(h.sub.nm)
[0200] Embodiments of the present disclosure provide a method that
the codebook for X is subset of
{ [ 1 j2.pi. n 16 ] , n = 0 , 1 , , 15 } , ##EQU00188##
where the subset comprises 8 distinct vectors indexed approximately
same resolution for UE 116 located within angular spread around 90
degree as O-bit codebook proposed in REF 3. As compared with Error!
Reference source not found. (evenly oversampled DFT vectors), beam
vectors in near 0 and 180 degrees (vectors represented by
{ [ 1 j2.pi. n 16 ] , n = 5 , 10 } ) ##EQU00189##
with lower density. If UEs are densely located within small angular
spread centering at 90 degree, our 3-bit codebook yields a similar
performance as the evenly oversampled 4-bit 4-Tx DFT codebook.
[0201] Although the present disclosure has been described with an
exemplary embodiment, various changes and modifications may be
suggested to one skilled in the art. It is intended that the
present disclosure encompass such changes and modifications as fall
within the scope of the appended claims.
* * * * *